September 2010 - Boost-users - lists.preview.boost.org

Boost.Random and Dieharder et al.
by Patricia Hawley 02 Sep '10

02 Sep '10

Hi, I am working with a team developing some utilities for a simulation, and one of our requirements is to test the quality of the random generators. We based the utility on Boost.Random, but now have to show that that was a good decision. Has anyone verified the generators (e.g. Mersenne Twister or Linear Congruential ) in Boost.Random pass the Dieharder, Ent, and/or TestU01 suites? Any comments or suggestions on this are appreciated, Pat

2 1

Re: [Boost-users] Boost-users Digest, Vol 2466, Issue 3
by Clarke, Carus V 01 Sep '10

01 Sep '10

Steven Watanabe wrote: > AMDG > > Clarke, Carus V wrote: > >> It looks like stat is setting errno to ENOENT. > >> Do you get the same behavior if the file exists before > starting the > >> program? > > > > Yes, the behavior is the same whether the program writes > the file or if the file had been previously written. > > > > Okay. Try using stat directly, instead of using Boost.Filesystem. > If that works, then can you use a debugger to find exactly > what arguments Boost.Filesystem is passing to stat? > > In Christ, > Steven Watanabe > Actually, the Boost.Filesystem library for Boost 1.44 uses the Windows API GetFileAttributes functions for reading file status on Cygwin, and it was in calls to these functions that things were failing. Since my last post, the Cygwin developers have updated the base Cygwin system (to 1.7.7-1). Updating my Cygwin installation to the new version has fixed the problem. Apparently, the 1.7.6-1 version had some bugs that were inadvertently introduced in an attempt to improve Cygwin functionality. I do have a question that might be better asked in a new thread, but here it is. Why does the Boost.Filesystem library force the use of Windows-style paths for Cygwin? The Cygwin system prefers Posix or mixed paths with forward slashes as separators. The use of Win32 paths on Cygwin is now deprecated, and Cygwin actually issues a warning the first time it encounters a Windows-style path. See http://cygwin.com/cygwin-ug-net/using.html#using-pathnames for more information. The Cygwin GCC 4.3.4 compiler no longer supports the -mno-cygwin option that was supported under Cygwin 1.5 and earlier. This means that any application created with this compiler depends on the Cygwin dll. Also, Boost does not appear to have a Cygwin platform for regression testing. Would it help if I or someone else could set one up? Thanks for your help. Carus V. (Bud) Clarke carus.v.clarke(a)boeing.com

1 0

Problem using serialization::xml_archive with polymrphic classes
by przemyslaw.sliwa＠uk.bnpparibas.com 01 Sep '10

01 Sep '10

HelloHello, I have written code below and it works fine when using text archives. However, when I try to run it with the xml archives it is throwing a strange error saying "Invalid XML tag name". I have no idea what that means and could not find it in the documentation of the serialization library. Could someone tell me what I am doing wrong in this code? Thanks a lot for help! #include <boost/smart_ptr/shared_ptr.hpp> #include <boost/archive/xml_oarchive.hpp> #include <boost/archive/xml_iarchive.hpp> #include <boost/archive/text_oarchive.hpp> #include <boost/archive/text_iarchive.hpp> #include <boost/serialization/vector.hpp> #include <boost/serialization/string.hpp> #include <boost/serialization/shared_ptr.hpp> #include <boost/date_time/local_time/local_time.hpp> #include <boost/serialization/export.hpp> #include <iostream> using namespace boost; using namespace std; class Actor { public: Actor( void ) {} Actor( char gender ) : gender( gender ) {} virtual ~Actor() {} char Gender( void ) const { return gender; } private: friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int version) { ar & BOOST_SERIALIZATION_NVP(gender); } char gender; }; class Crit: public Actor { public: Crit( void ) {} Crit( int age ) : age( age ) { } int Age( void ) const { return age; } private: friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int version) { ar & BOOST_SERIALIZATION_NVP(boost::serialization::base_object< Actor >( *this )); ar & BOOST_SERIALIZATION_NVP(age); } int age; }; class Food: public Actor { public: Food( void ) {} Food( int energy ) : energy( energy ) { } int Energy( void ) const { return energy; } private: friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int version) { ar & BOOST_SERIALIZATION_NVP(boost::serialization::base_object< Actor >( *this )); ar & BOOST_SERIALIZATION_NVP(energy); } int energy; }; BOOST_CLASS_EXPORT_GUID( Actor, "Actor") BOOST_CLASS_EXPORT_GUID( Crit, "Crit") BOOST_CLASS_EXPORT_GUID( Food, "Food") template< typename T > void save( T& data, string filename ) { ofstream ofs( filename.c_str() ); boost::archive::xml_oarchive oa( ofs ); oa & BOOST_SERIALIZATION_NVP(data); } template< typename T > void load( T& data, string filename ) { ifstream ofs( filename.c_str() ); boost::archive::xml_iarchive ia( ofs ); ia & BOOST_SERIALIZATION_NVP(data); } void TestAlpha( void ) { shared_ptr< Crit > critAlpha( new Crit( 18 ) ); shared_ptr< Crit > critBravo( new Crit( 19 ) ); shared_ptr< Food > foodAlpha( new Food( 20 ) ); shared_ptr< Food > foodBravo( new Food( 21 ) ); vector< shared_ptr< Actor > > conAlpha; vector< shared_ptr< Actor > > conBravo; conAlpha.push_back( critAlpha ); conAlpha.push_back( foodAlpha ); // save string filename = "sav.sav"; save( conAlpha, filename ); // test assert( critBravo->Age() == 19 ); assert( foodBravo->Energy() == 21 ); // load load( conBravo, filename ); // test critBravo = dynamic_pointer_cast< Crit >( conBravo[ 0 ] ); assert( critBravo->Age() == 18 ); foodBravo = dynamic_pointer_cast< Food >( conBravo[ 1 ] ); assert( foodBravo->Energy() == 20 ); } int main() { try { TestAlpha(); } catch(boost::archive::archive_exception& e) { return -1; } return 0; } ___________________________________________________________ This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and delete this e-mail. Any unauthorised copying, disclosure or distribution of the material in this e-mail is prohibited. Please refer to http://www.bnpparibas.co.uk/en/information/legal_information.asp?Code=ECAS-… for additional disclosures.

2 1

mpl::set and mpl::fold
by Paul Graphov 01 Sep '10

01 Sep '10

Hello Boost Users, I've encountered a problem while using mpl::fold with mpl::set I've defined a template to check if one sequence contains all elements of the other: namespace mpl = boost::mpl; template<class A, class B> struct subset : mpl::fold< A, mpl::true_, mpl::and_<mpl::_1, mpl::contains<B, mpl::_2> > > {}; It works fine with lists and vectors but when first template parameter (A) is set or set_c it fails to compile. Did I miss something? Why is it impossible to iterate through a set via fold? p.s. tested on gcc and clang under linux. p.p.s. compilation error is quite huge, the main message is /usr/local/include/boost/mpl/eval_if.hpp:38: error: no type named ‘type’ in ‘struct boost::mpl::next<boost::mpl::s_iter<boost::mpl::set_c<int, 1l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l, 9223372036854775807l>, boost::mpl::set0_c<int> > >’

2 3

[gil] bit_aligned_pixel_iterator
by Christian Henning 01 Sep '10

01 Sep '10

Hi there, I have a problem with extracting the correct values out of a 3bit gray image. The test code below displays my problem. Here, there is a vector of unsigned chars which create a particular bit pattern though all pixels in the image should have the same values. The bit pattern is: 1011 0110 0110 1101 1101 1011 0000 0000. On my machine ( x86 ) this should be read out in values of 110b ( which is 6 ) for all pixels. Instead my code reads out: 6, 6, 2, 6, 6, 0, 6, 6, 0. Here is the code: #include <cassert> #include <vector> #include <boost/gil/gil_all.hpp> using namespace std; using namespace boost; using namespace gil; int main() { typedef bit_aligned_image1_type< 3, gray_layout_t >::type gray3_image_t; typedef gray3_image_t image_t; typedef image_t::view_t view_t; typedef view_t::reference ref_t; typedef view_t::value_type pixel_t; typedef bit_aligned_pixel_iterator< ref_t > iterator_t; vector< unsigned char > buf( 4 ); buf[0] = 182; buf[1] = 109; buf[2] = 219; buf[3] = 0; iterator_t it( &buf[0], 0 ); ref_t p1 = *it; it++; ref_t p2 = *it; it++; ref_t p3 = *it; it++; ref_t p4 = *it; it++; ref_t p5 = *it; it++; ref_t p6 = *it; it++; ref_t p7 = *it; it++; ref_t p8 = *it; it++; ref_t p9 = *it; it++; unsigned char v1 = get_color( p1, gray_color_t() ); unsigned char v2 = get_color( p2, gray_color_t() ); unsigned char v3 = get_color( p3, gray_color_t() ); unsigned char v4 = get_color( p4, gray_color_t() ); unsigned char v5 = get_color( p5, gray_color_t() ); unsigned char v6 = get_color( p6, gray_color_t() ); unsigned char v7 = get_color( p7, gray_color_t() ); unsigned char v8 = get_color( p8, gray_color_t() ); unsigned char v9 = get_color( p9, gray_color_t() ); return 0; } The first two values are correct since they covered by the first byte. The third value doesn't seem to take in account the first bit of the second byte. I think this is incorrect. The stepping of the bits of by the iterator seems correct. Each value that is covered by one byte completely is correct but when a value is stretched over two bytes there seems to be a problem. Regards, Christian

1 1

[type_traits]: is_boost_optional?
by Roland Bock 01 Sep '10

01 Sep '10

Hi, is there something in Boost.TypeTraits (or somewhere else) that allows me to check if a type is a boost::optional? I'd like to do something like this: template<typename T> typename enable_if<is_boost_optional<T> >, T::value_type>::type foo(const T&); Thanks and regards, Roland

3 7

Boost.Filesystem: question about code conversion for char16_t with version 3
by Yasuhiro Hara 01 Sep '10

01 Sep '10

Hello, I am studying Boost.Filesystem version 3, but I got a couple of questions. I am wondering if someone helps me to understand version 3. I saw the reference document and path.hpp header file, I am not sure how to specify a code conversion API for char16_t or char32_t. It looks like only code conversion API between wchar_t and char can be specified. Since version 3 tries to support char16_t and char32_t, I am a little puzzled. I would appreciate if someone helps me to understand how I can use char16_t with version 3. Best regrads, /yasu

1 0

compressed fusion::vector?
by Cory Nelson 01 Sep '10

01 Sep '10

Hello, Is there a compressed fusion::vector that minimizes the size if the types passed in are empty, like compressed_pair? Thanks! -- Cory Nelson http://int64.org

2 1

[iostreams] filtering_ostream does not flush std::cout.
by Duncan Smith 01 Sep '10

01 Sep '10

I'm using Boost 1.43.0, and I've found that filtering_ostream does not flush "std::cout". I found a relevant (and recent) ticket against Boost 1.44.0, but my issue is slightly different. https://svn.boost.org/trac/boost/ticket/4590 I'm using GCC 4.3.4 on Gentoo Linux. I'm hoping that I'm just doing something wrong, but if you think there's a bug, I'll add my testcase to the ticket. The basic recipe: filtering_ostream out; out.push(any_filter_at_all()); out.push(std::cout); out << "Hello World!" << std::flush; sleep(10); // Observe an empty console. std::cout << std::flush; sleep(10); // Observe printed text on console. This problem is quite bad for me when doing the following on the command-line (for a more complicated program): ./program 2>&1 | tee some-file.log Because stdout is not getting flushed, the stdout and stderr get intermingled strangely. However, disabling buffering on stdout works around the problem: stdbuf --output=0 ./program 2>&1 | tee some-file.log Please tell me if you think there's a bug in my code. If it's a bug in Boost.Iostreams (or I'm expecting the wrong behaviour), I would appreciate any ideas for workarounds (I tried cout.setf(unitbuf), and that doesn't seem to do the trick). Here's the full text of my code: ------------------ // Test case for flushing a filtering_ostream based on the // boost_iostreams_filtering_ostream.cpp attached to ticket #4950: // https://svn.boost.org/trac/boost/ticket/4590 // https://svn.boost.org/trac/boost/attachment/ticket/4590/boost_iostreams_fil… // changed to use "cout" as the sink. // // Flushing a filtering_ostream using the flush() global function does not work // in Boost 1.43. I suspect it also does not work in Boost 1.44. // // Using GCC-4.3.4 on Gentoo Linux. // #include <cassert> #include <string> #include <boost/iostreams/categories.hpp> #include <boost/iostreams/operations.hpp> #include <boost/iostreams/filtering_stream.hpp> #include <iostream> // Included for "sleep()" call below. #include <unistd.h> namespace io = boost::iostreams; // Simple output filter forwarding just forwarding the put chars to // the sink. struct transparent_output_filter { typedef char char_type; typedef io::output_filter_tag category; template <typename Sink> bool put(Sink& snk, char c) { return io::put(snk, c); } }; int main(int argc, char *argv[]) { using namespace std; io::filtering_stream<io::output> out; // Transparent filter just forward the characters to the sink. // // Note: I observe this behaviour with and without this line. out.push(transparent_output_filter()); // Sink appends characters to "cout". out.push(cout); // Write test string to filtering_ostream. out << "Hello World!"; // On my system, this will flush the buffer iff I'm on an // interactive console. "\n" and endl have the same effect. I believe // that my shell sets "stdout" to be line-buffered if it's interactive. // out << "\n"; bool wasFlushed = false; // The flush does *not* work in Boost 1.43. wasFlushed = io::flush(out); // This assertion does not go off, though. assert(wasFlushed); // This also doesn't work. wasFlushed = out.flush(); assert(wasFlushed); // This also doesn't work. out << std::flush; // Demonstrate that none of the above flush calls had any effect on "cout". sleep(10); // Now we finally get the characters on the console. cout << flush; // Demonstrate that the direct flush call works when called directly on // "cout". sleep(10); return 0; }

1 0

Can I use both filesystem versions v2 and v3 in a program ?
by Timothy Madden 01 Sep '10

01 Sep '10

Hello Can I include <boost/filesystem.hpp> but use the BOOST_FILESYSTEM_V... macro to include one version in on of my cpps and another version in another cpp ? Do both versions get compiled when boost is built ? Thank you, Timothy Madden

1 0

Re: [Boost-users] Boost.Serialization: Restoringobjectstate(preventcreation of a new object)?
by ray＠maya-studio.com 01 Sep '10

01 Sep '10

I've ended up with the following solution: #include <fstream> #include <iostream> #include <boost/archive/xml_iarchive.hpp> #include <boost/archive/xml_oarchive.hpp> #include <boost/archive/text_iarchive.hpp> #include <boost/archive/text_oarchive.hpp> #include <boost/serialization/export.hpp> #include <boost/foreach.hpp> #define _SERIALIZATION_FUNC(CLASS_NAME, ARCH_TYPE) \ public: \ virtual void Serialize(ARCH_TYPE& ar) \ { \ CLASS_NAME& x = *static_cast<CLASS_NAME*>(this); \ ar & BOOST_SERIALIZATION_NVP(x); \ } #define SERIALIZATION(CLASS_NAME) \ public: \ _SERIALIZATION_FUNC(CLASS_NAME, boost::archive::text_iarchive) \ _SERIALIZATION_FUNC(CLASS_NAME, boost::archive::text_oarchive) \ _SERIALIZATION_FUNC(CLASS_NAME, boost::archive::xml_iarchive) \ _SERIALIZATION_FUNC(CLASS_NAME, boost::archive::xml_oarchive) \ private: \ friend class boost::serialization::access; \ template<typename TArchive> \ void serialize(TArchive & ar, const unsigned int version /*file_version*/) class A { public: A() : m_a(0) {} protected: int m_a; private: SERIALIZATION(A) { ar & BOOST_SERIALIZATION_NVP(m_a); } }; class B : public A { public: B() : m_b(0) {} protected: int m_b; private: SERIALIZATION(B) { ar & BOOST_SERIALIZATION_BASE_OBJECT_NVP(A); ar & BOOST_SERIALIZATION_NVP(m_b); } }; BOOST_CLASS_EXPORT(B) int main(int argc, char* argv[]) { // ... A a; B b; std::list<A*> lst; // List of A* pointers lst.push_back(&a); // Base class lst.push_back(&b); // Derived //... // Save { std::ofstream ofs("c:/test.xml"); boost::archive::xml_oarchive oa(ofs); BOOST_FOREACH(A* pA, lst) { pA->Serialize(oa); } } //... // Restore { std::ifstream is("c:/test.xml"); boost::archive::xml_iarchive ia(is); BOOST_FOREACH(A* pA, lst ) { pA->Serialize(ia); } } //... } Everything works as expected (my classes can live with a few additional virtual functions :). And thank you for creating this excellent serialization library! Ray

1 0

Using Boost.MPI with gather , data becomes overwritten after gather operation
by hgc-01134＠hkedcity.net wong 01 Sep '10

01 Sep '10

Hi, I am trying to use Boost.MPI to do a very simple program that 1. Processes from different ranks do a gather operation 2. The gather operation will gather the CoProcessInfo ( which is just a data structure of rank# and hostname) into a vector 3. After the gather operation, Rank 0 will print the CoProcessInfo one by one in the vector. ------------------------------------------------------------------------------------------------------------------------------------------------------ My testing environment consists of TWO virtual machine ( each of the machine has two cores ): Centos 5.5 mpich2 Boost.mpich2 Boost 1.41 mpiexec ------------------------------------------------------------------------------------------------------------------------------------------------------ The expected output of the program should be something like ( just an example order of ranks ): Host VM1, Rank 0 Host VM1, Rank 1 Host VM2, Rank 2 Host VM2, Rank 3 ---------------------------------------------------------------------------------------------------------------------------------------------------------- However, I got something like: ( just an example order of ranks ), the hostname is incorrect after gather operation ( three VM2 appears!) Host VM1, Rank 0 Host VM2, Rank 1 Host VM2, Rank 2 Host VM2, Rank 3 I found that each process CAN correctly found out the hostname and rank#, however, the data will become a bit like overwritten, after the gather operation. Thank you very much. ------------------------------------------- Source Code ------------------------------------------------------------------- #include <boost/asio/ip/host_name.hpp> #include <iostream> #include <vector> #include <string> #include <boost/mpi/environment.hpp> #include <boost/mpi/communicator.hpp> #include <boost/mpi.hpp> #include <boost/archive/text_oarchive.hpp> #include <boost/archive/text_iarchive.hpp> #include <sstream> #include <fstream> using namespace std; namespace mpi = boost::mpi; class CoProcessInfo { private: string hostname; int rank; friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int /* version */) { ar & hostname & rank; } public: CoProcessInfo() { hostname = boost::asio::ip::host_name(); rank = 0; } void setRank(int _rank) { rank = _rank; } string getHostname() const { return hostname; } int getRank() const { return rank; } }; int main(int argc, char* argv[]) { int myrank; mpi::environment env(argc, argv); mpi::communicator world; myrank = world.rank(); CoProcessInfo localInfo; localInfo.setRank( myrank ); std::vector<CoProcessInfo> coProcInfo; mpi::gather(world, localInfo, coProcInfo, 0); if( myrank == 0 ){ for (std::vector<CoProcessInfo>::const_iterator it = coProcInfo.begin(); it !=coProcInfo.end(); ++it) { cout << endl << " Host : " << it->getHostname() << " , Rank : " << it->getRank() << endl; } } return 0; } ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Regards, Joseph

2 1