June 2013 - Boost - lists.boost.org

Tips to fix boost::asio::windows::random_access_handle which is seriously broken
by Niall Douglas 13 Jun '13

13 Jun '13

Dear Boost, The forthcoming GSoC Boost.AFIO library has found a problem with boost::asio::windows::random_access_handle in that it appears to clamp buffer sizes to 64Kb and not increment overlapped offsets during multi-buffer reads and writes, and which is reported at https://svn.boost.org/trac/boost/ticket/8669. These are desirable features for network i/o, not so for file i/o. As my work on the existing code base which Paul is porting to Boost is now complete and purring beautifully on POSIX, I'm thinking I might as well go fix the Boost.ASIO Windows IOCP problem. Before I begin, can I ask here if anyone else has done some work on this topic from which I might be able to borrow? The thread at http://comments.gmane.org/gmane.comp.lib.boost.asio.user/3676 suggests that if I use async_write_some() then I can bypass the 64Kb clamp using "my own completion condition". I'm thinking that boost::asio::windows::random_access_handle really ought to actually implement random access, so rather than hacking around the problem, fixing boost::asio::windows::random_access_handle is better. My thanks in advance, Niall --- Opinions expressed here are my own and do not necessarily represent those of BlackBerry Inc.

1 0

[Multiprecision] High-performance radix-2 floating-point back-end
by Aditya Jain 13 Jun '13

13 Jun '13

Hi This is my first post on the mailing list , so apologies if it doesn't follow any of the protocols. I was looking at GSoC projects this year for Boost and I am wondering if someone is working on a high-performance radix-2 floating-point back-end for Boost.Multiprecision. If not , I would be interested in doing this project .I am currently a Masters student interning at a tech company in Seattle but I would be able to devote 15-20 hours a week for this project. I have a fairly good background in C++ and numerical programming.So in case no else is working on it (for GSoC or otherwise ) , I would be really interested . If someone is already working on it, I would be interested in some other projects I could potentially work on . Looking forward to a reply. Regards -- Aditya Jain MS Computer Science (Machine Learning Track) Columbia University B.Tech, IIT Roorkee

4 4

ATTN: Boost GSoC mentors and students: Continuous Integration for GSoC thoughts
by Niall Douglas 12 Jun '13

12 Jun '13

Dear Boost, I spent the weekend experimenting with getting CI set up for the async file i/o library which will hopefully become Boost.AFIO through Paul's GSoC. I tried a number of free-for-open-source CI providers which integrate tightly with GitHub, and eventually settled with Travis CI. You can see the last commit CI results at https://travis-ci.org/ned14/TripleGit. As much as Travis CI has a primitive and clunky UI, unlike the other CIs it's based on snapshotting VMs with root privs which gives you unparalleled flexibility for testing config, so I was able to configure this to happen with each and every git push to github: 1. Test that code builds with GCC 4.6 and clang 3.2. 2. Test that unit and functional tests execute without error when built with GCC 4.6 and clang 3.2. 3. Run unit and functional tests with valgrind --tool=memcheck and test that no memory errors nor memory leaks occur with binaries built with GCC 4.6 and clang 3.2. 4. Install GCC 4.8 and compile everything using -fsanitize=thread (i.e. instrument using the port of clang's ThreadSanitizer v2 to GCC), then execute a full set of unit and functional tests looking for race conditions and if any, fail. You'll note I've marked that test as allowed to fail, because Boost v1.53 has a number of known race conditions when used as C++11 which are hopefully fixed in v1.54 :) Personally speaking, I think this depth and degree of CI is *invaluable* to modern software development practice. So the question becomes: are the Boost GSoC mentors and students interested in setting up CI for this year's GSoC projects, or at least the ones hosted on github at https://github.com/BoostGSOC? In case you're wondering how hard it is to set up Travis CI for your github project, you just need one file called .travis.yml. You can see the one I wrote at https://github.com/ned14/TripleGit/blob/master/.travis.yml. Thoughts? Niall --- Opinions expressed here are my own and do not necessarily represent those of BlackBerry Inc.

4 7

[utility] [assert] Proposal for a new assert macro which stores the failed values ("informative assert")
by Viktor Sehr 11 Jun '13

11 Jun '13

I'd like some opinions about this assert implementation which I'd like to propose as an update of the old BOOST_ASSERT. Basically it stores the failed values next to the expression string, instead of just the expression string: "INFORMATIVE_ASSERT(5*5 < 2*2);" "5*5 < 2*2 | 25 | 4" "INFORMATIVE_ASSERT(std::vector<int>(10).size() == std::vector<int>(12).size());" "std::vector<int>(10).size() > std::vector<int>(12).size() | 10 | 12" Notes - If the class can't be serialized, the byte-data is logged instead. - It writes the values to the expression string in order to be compliant with any assert implementations Easy to read implementation available here: https://docs.google.com/file/d/0B69BCPzak52DcnF5dmhRRno3V2M/edit?usp=sharing Modified version of BOOST_ASSERT available here: https://docs.google.com/file/d/0B69BCPzak52DZW1oX1AtREM4eVU/edit?usp=sharing /Viktor (Reposting this as I posted it without labels one week ago)

4 13

Re: [boost] [1.54.0] Beta 1 Available
by Cromwell Enage 11 Jun '13

11 Jun '13

Hi, all. When calling the merge() member function of a boost::heap::d_ary_heap, is the heap passed in not supposed to be emptied? Here's my test program: ['' #include <boost/heap/d_ary_heap.hpp> #include <boost/heap/binomial_heap.hpp> #include <boost/heap/fibonacci_heap.hpp> #include <boost/heap/pairing_heap.hpp> #include <boost/heap/skew_heap.hpp> #include <boost/test/minimal.hpp> template <typename Heap> void test_heap_merge_memfun() { Heap heap_1, heap_2; heap_1.push(3); heap_1.push(5); heap_1.push(1); heap_2.push(2); heap_2.push(4); heap_2.push(0); heap_1.merge(heap_2); std::cout << " heap_2"; if (!heap_2.empty()) { BOOST_CHECK(4 == heap_2.top()); heap_2.pop(); BOOST_CHECK(2 == heap_2.top()); heap_2.pop(); BOOST_CHECK(0 == heap_2.top()); heap_2.pop(); std::cout << " not"; } std::cout << " emptied." << std::endl; BOOST_CHECK(heap_2.empty()); BOOST_CHECK(5 == heap_1.top()); heap_1.pop(); BOOST_CHECK(4 == heap_1.top()); heap_1.pop(); BOOST_CHECK(3 == heap_1.top()); heap_1.pop(); BOOST_CHECK(2 == heap_1.top()); heap_1.pop(); BOOST_CHECK(1 == heap_1.top()); heap_1.pop(); BOOST_CHECK(0 == heap_1.top()); heap_1.pop(); BOOST_CHECK(heap_1.empty()); } int test_main(int argc, char** argv) { test_heap_merge_memfun< boost::heap::d_ary_heap< int , boost::heap::arity<4> , boost::heap::mutable_<true> > >(); test_heap_merge_memfun<boost::heap::binomial_heap<int> >(); test_heap_merge_memfun<boost::heap::fibonacci_heap<int> >(); test_heap_merge_memfun<boost::heap::pairing_heap<int> >(); test_heap_merge_memfun< boost::heap::skew_heap<int,boost::heap::mutable_<true> > >(); return 0; } ''] When I run this program, I see that all the other heap types empty the heap passed in. I use gcc 4.2.1 with Apple clang version 4.0 on MacOSX 10.8 Cromwell D. Enage

2 1

[parameter] BOOST_PARAMETER_CONSTRUCTOR bug
by James Hirschorn 10 Jun '13

10 Jun '13

I've already submitted a bug report, but I'm also posting it here in case anyone has more info. ------------- The following variation on the example from the boost::parameter tutorial, where the required parameter is changed to optional, fails to compile. myclass_impl is being passed an empty argument list. #include <iostream> #include <boost/parameter.hpp> BOOST_PARAMETER_NAME(name) BOOST_PARAMETER_NAME(index) struct myclass_impl { template <class ArgumentPack> myclass_impl(ArgumentPack const& args) { std::cout << "name = " << args[_name] << "; index = " << args[_index] << std::endl; } }; struct myclass : myclass_impl { BOOST_PARAMETER_CONSTRUCTOR( myclass, (myclass_impl), tag , (optional (name, *) (index, *) ) ) // no semicolon }; int main(int argc, char * argv[]) { myclass C(_name="hello", _index=0); }

3 6

[type_erasure] Compilation time
by Felipe Magno de Almeida 10 Jun '13

10 Jun '13

Hello Steven, Is there anyway to not include some definitions in header to delay instantiation for use with explicit template instantiations in specific translation units with type_erasure? I have a CORBA library which uses type_erasure for each interface defined in IDL. The problem is that if I include too many #include's for the precompiled headers which use type_erasure, then the compiler ends up consuming too much memory (>10GB), which makes it unfit for 32bit compilers which can use at most 3GB. Regards, -- Felipe Magno de Almeida

2 2

Re: [boost] [multiprecision] Are these types suitable as numeric types for unit tests?
by Daryle Walker 09 Jun '13

09 Jun '13

-----Original Message----- From: John Maddock Sent: Saturday, June 08, 2013 4:40 AM >> I'm getting warnings about integer overflow. The problem is at Line >> 59 of boost/test/tools/detail/print_helper.hpp. It sets an output >> stream's precision based on the digits given by std::numeric_limits. > >That's harmless but annoying: it's because std::numeric_limits<cpp_int>::digits is >INT_MAX. The stream precision isn't ever used by integer types anyway... I switched to one of the fixed-sized types (512 bits). >> This time it's an error, at Line 59 of >> boost/test/tools/floating_point_comparison.hpp. There no (close >> enough) matching function for "fpt_abs" from some >> "boost::multiprecision::detail::expression<..." type. (It was >> truncated by the display. I don't know how to get a clean print-out >> of errors from >> Code::Blocks.) I'm guessing the expression template setup is screwing >> things up. > >Yes, it's because Boost.Test can't be used with expression template enabled types. >It's a known issue, and I've added a patch for this: >https://svn.boost.org/trac/boost/ticket/8272, but Genediiy seems to be MIA at present :-( > >Do you particularly want to test with expression-template enabled types (I reaslise >there are cases when you do)? If not then you could use multiprecision types that >have expression templates turned off and that should fix things. I switched to the same type with expression-templates turned off. Daryle W.

1 0

[range] [extensions] Default values for range access adaptors?
by Viktor Sehr 09 Jun '13

09 Jun '13

I'd like to suggest the addition of a default value to the range access adaptors. Either by adding it as an extra, non-necessary parameter for the functions, or by duplicating the functions as "front_or_default(T default value)". For example it would be useful for finding objects which I am not sure exists in a range. std::vector<int> vec = boost::assign::list_of(2)(4)(6); auto first_odd = vec | filtered(is_odd) | front(); // undefined? auto first_odd = vec | filtered(is_odd) | front_or_default(-1); // everything is fine! Another option would be a function which returns the adress, or nullptr if the value does not exist? /Viktor

5 14

boost.range bug in transformed or sliced
by John Reid 09 Jun '13

09 Jun '13

Suppose you wish to transform and slice a random access range. Surprisingly the order in which you apply the operations makes a big difference to performance. I'm guessing this is not desired behaviour, correct me if I'm wrong. AFAICT when transforming happens before slicing the iterators involved in the internals of boost.range are being treated as forward iterators not random access iterators. An advance() operation is made on one of the iterators with a negative number and this advance is O(n) for forward iterators but O(1) for random access. The code below demonstrates (adapted from sliced example code). On an unrelated note the documentation for transformed does not mention that the function is part of the range return type. It is given as: boost::transformed_range<typeof(rng)> Code to demonstrate the transform then slice problem: #include <boost/range/adaptor/transformed.hpp> #include <boost/range/adaptor/sliced.hpp> #include <boost/range/algorithm/copy.hpp> #include <boost/assign.hpp> #include <iterator> #include <iostream> #include <vector> #include <functional> struct identity { typedef int result_type; result_type operator()( int i ) const { return i; } }; int main(int argc, const char* argv[]) { using namespace boost::adaptors; using namespace boost::assign; std::vector< int > input; input += 1,2,3,4,5,6,7,8,9; // slicing then transforming my iterator range std::cout << "Sliced then transformed: "; boost::copy( input | sliced( 2, 8 ) | transformed( identity() ), std::ostream_iterator< int >( std::cout, ",") ); std::cout << "\n"; // transforming then slicing my iterator range - takes a very long time.... std::cout << "Transformed then sliced: "; boost::copy( input | transformed( identity() ) | sliced( 2, 8 ), std::ostream_iterator< int >(std::cout, ",")); std::cout << "\n"; return 0; }

4 5