- Boost - lists.boost.org

[1.45][filesystem/io] Streaming operator
by Chard 12 Nov '10

12 Nov '10

The filesystem path streaming operator (>>) has a bug in 1.44, documented here: https://svn.boost.org/trac/boost/ticket/4716 A patch has been provided, and has been implemented on the trunk, detailed here: http://lists.boost.org/boost-commit/2010/08/27576.php Any chance of this being merged to the release branch for 1.45? [.../boost/io/detail/quoted_manip.hpp] --

1 0

Re: [boost] [Boost-users] What's so cool about Boost.MPI?
by Matthias Troyer 12 Nov '10

12 Nov '10

On 5 Nov 2010, at 20:04, Riccardo Murri wrote: > On Fri, Nov 5, 2010 at 3:51 PM, Peter Foelsche <foelsche(a)sbcglobal.net> wrote: >> Can somebody explain to me, what is so great about mpi or boost/mpi? > > Daunting task, but I'll give it a try :-) > > MPI is a library interface[1] to ease some common communication > patterns in scientific/HPC software. In particular, MPI provides > library functions for: > > 1. Reliable delivery of messages between two concurrent processes; > there are functions for both synchronous and asynchronous > communication. > > 2. Collective communication of the running processes: e.g., one > process broadcasting a message to all others; all processes > broadcasting to all; scattering a large array to all running > processes, etc. > > 3. "Remote shared memory": a process can expose a memory buffer that > all others can directly read/write. > > 4. Parallel I/O, where many processes access (possibly interleaved) > portions of a file independently. > > Despite being non-trivial features to implement on top of a standard > TCP socket interface, MPI implementations are mostly into performance: > they can usually take advantage of special hardware (e.g., Infiniband > or Myrinet interconnects) and use the native protocols (instead of > TCP/IP) to provide faster communication speed. > > A "message" in usual socket programming is just a flat array of bytes. > A "message" in MPI is a (potentially non-contiguous portion) of a C or > FORTRAN data structure; MPI can handle C structures and > multi-dimensional C-style arrays, and multi-dimensional arrays of C > structures (containing multi-dimensional arrays, etc.). However, > since MPI is implemented as a library, you have to tdescribe to MPI > the in-memory layout of the data structures you want to send as a > message; thus "duplicating" work that is done by the compiler. You > can imagine that this quickly gets tedious and unwieldy once you start > having non-trivial data structures, possibly with members of unknown > length (e.g., a list). > > This is where Boost.MPI comes to the rescue: if you can provide > Boost.Serialization support for your classes, you can send them as MPI > messages with no extra effort: no need for lengthy calls to > MPI_Type_create_struct() etc. And it does this with minimal or even > no overhead! (Plus you also get a nice C++ interface to MPI > functionality, whereas the MPI C calls are quite low-level.) > > > .. [1] MPI proper is a specification; there are several > implementations of the spec (e.g., OpenMPI, MPICH, MVAPICH, plus > many vendor/proprietary one), but they are all compatible at the > source level. > > Best regards, > Riccardo Nice summary! Matthias

5 8

[1.45.0] Beta 1 available from SourceForge
by Beman Dawes 11 Nov '10

11 Nov '10

Boost release1.45.0 beta 1 is now available from SourceForge. For those who downloaded release candidate 1, note that beta 1 was build from a slightly later release candidate. See http://sourceforge.net/projects/boost/files/ This release contains numerous enhancements and bug fixes for existing libraries. For details of what's in the release, see http://beta.boost.org/users/news/version_1_45_0. Note that the links to files on this web page are for the final release - use the SourceForge link above to get the beta files. Please download the beta, give it a try, and report any problems you encounter. There are still a few serious bugs outstanding, and we will be working on clearing those before the final release. Thanks, -- The Boost Release Team

7 9

Re: [boost] [serialization][thread] possible regressions?
by Eric Niebler 11 Nov '10

11 Nov '10

On 11/11/2010 1:41 PM, Robert Ramey wrote: > Anthony Williams wrote: >> Eric Niebler <eric(a)boostpro.com> writes: >> >>> I don't know whether this is a serialization issue or a thread issue. >>> Robert, Anthony, can you two weigh in? >> >> Looks like a serialization problem to me. > > I would agree - but its hard to reproduce from here. I don't see this as > quick fix - it would require some time and help from the person who can > actually reproduce it. Thanks, Robert. This is now https://svn.boost.org/trac/boost/ticket/4842 Ruediger, anything you can do to help Robert with this one would be much appreciated. -- Eric Niebler BoostPro Computing http://www.boostpro.com

1 0

Problem with boost deserialization.
by Esteban RG 11 Nov '10

11 Nov '10

Hi, I need serialization in a project of mine. I have reduced my options to boost and ProtocolBuffer, but in order to be able to decide which one to use, I am trying to build a timing test for serialization and deserialization. Let me explain a little about my setup: a) Platform: - windows 7, 64bit version - visual studio 2008 express edition x32 b) Additional library: - ACE library for High resolution timer. The code (an explanations) that I have used to implement the comparison is at follows: I build a class with four different lists of boost/serialization/list like documentation states: std::list<unsigned int> &lint_; std::list<ACE_INT64> &llong_; std::list<float> &lfloat_; std::list<std::string> &lstring_; The test is increasing the size of every list simultaneously. Then it gets serializated and deserializated and finally I get the time results for this process. However there is one issue: Serialization always works, but deserialization fails during execution time whenever the lists are bigger than 10 elements, throws a exception "input stream error" into boost/serialization/throw_exception.hpp as seen with the debugger in the variable/const std::exception const &e (boost library) Visual studio shows this message: Unhandled exception at 0x765cb727 in prTestBoostSerialization2.exe: Microsoft C++ exception: boost::archive::archive_exception at memory location 0x003fc564.. my class definition is: #include <boost/archive/binary_oarchive.hpp> #include <boost/archive/binary_iarchive.hpp> #include <boost/serialization/string.hpp> #include <boost/serialization/list.hpp> //#include <boost/serialization/version.hpp> //#include <boost/serialization/split_member.hpp> #include <iostream> #include <list> #include <string> class BoostSerializationClass { public: // Not required because al member are public // friend class boost::serialization::access; BoostSerializationClass(std::list<unsigned int> &lint , std::list<float> &lfloat , std::list<ACE_INT64> &llong , std::list<std::string> &lstring) : lint_(lint), lfloat_(lfloat), llong_(llong), lstring_(lstring) { //std::cout << std::endl << " max size " << lint_.max_size(); //std::cout << std::endl << " " << llong_.max_size(); //std::cout << std::endl << " " << lfloat_.max_size(); //std::cout << std::endl << " " << lstring_.max_size(); } std::list<unsigned int> &lint_; std::list<ACE_INT64> &llong_; std::list<float> &lfloat_; std::list<std::string> &lstring_; }; namespace boost { namespace serialization { template<class Archive> void serialize(Archive &ar, BoostSerializationClass &c , const unsigned int version){ ar & c.lint_; ar & c.llong_; ar & c.lfloat_; ar & c.lstring_; } } } ----------------------- And main function is: #include <iostream> #include <fstream> //#include <cstdlib> #include <ace/OS.h> #include <ace/Time_Value.h> #include <ace/High_Res_Timer.h> // high resolution timer // include for BOOST #include "../prTestProtocolBuffers/BoostSerializationClass.h" #include <boost/archive/binary_oarchive.hpp> #include <boost/archive/binary_iarchive.hpp> //#include <list> <- dont include this use boost/serialization/list #include <boost/serialization/list.hpp> void gen_random_string(std::string &s, const int len); int ACE_TMAIN(int argc, ACE_TCHAR* argv[]){ //int lengthList = 5; //int lengthList = 10; int lengthList = 15; //int lengthList = 50; // int lengthList = 100; // int lengthList = 1000; std::list<unsigned int> listInteger,listInteger2; std::list<float> listFloat,listFloat2; std::list<ACE_INT64> listLongs,listLongs2; std::list<std::string> listString,listString2; int rint = 0; float rfloat = 0.f; ACE_INT64 rint64 = 0; std::string rstring = ""; std::cout << std::endl << "Begin random generate message. Please be patient."; // Generate random fields for messages for (int i=0 ; i<lengthList ; i++){ listInteger.push_back( rand() ); listFloat.push_back( (float)rand()/(float)RAND_MAX ); listLongs.push_back( rand()*rand()*rand() ); gen_random_string(rstring, lengthList); listString.push_back( rstring ); } // BOOST SERIALIZATION ACE_High_Res_Timer hrTimerBoost; // Load fields into BoostSerializableClass BoostSerializationClass boostSerializationClass( listInteger, listFloat, listLongs, listString); std::ofstream obfile("boost-file.bin"); boost::archive::binary_oarchive oba(obfile); //ACE_Time_Value btvStart = ACE_OS::gettimeofday(); hrTimerBoost.start(); oba & boostSerializationClass; // boost serialization; obfile.close(); hrTimerBoost.stop(); ACE_hrtime_t usecsBoost; hrTimerBoost.elapsed_microseconds(usecsBoost); //ACE_Time_Value btvEnd = ACE_OS::gettimeofday(); //ACE_Time_Value tvIncBoost( btvEnd.sec() - btvStart.sec() , btvEnd.msec() - btvStart.msec() ); std::cout << std:: endl << " ===== BOOST SERIALIZATION =====" ; //std::cout << std::endl << tvIncBoost.sec() << " seconds " << tvIncBoost.msec() << " mseconds" << std::endl ; std::cout << std::endl << usecsBoost << " micro seconds " << std::endl ; // BOOST DESERIALIZACION BoostSerializationClass boostSerializationClass2( listInteger2, listFloat2, listLongs2, listString2); std::ifstream ibfile("boost-file.bin"); boost::archive::binary_iarchive iba(ibfile); //btvStart = ACE_OS::gettimeofday(); hrTimerBoost.reset(); hrTimerBoost.start(); iba & boostSerializationClass2; ibfile.close(); hrTimerBoost.stop(); hrTimerBoost.elapsed_microseconds(usecsBoost); //btvEnd = ACE_OS::gettimeofday(); //tvIncBoost = ACE_Time_Value( btvEnd.sec() - btvStart.sec() , btvEnd.msec() - btvStart.msec() ); std::cout << std:: endl << " ===== BOOST DE-SERIALIZATION =====" ; //std::cout << std::endl << tvIncBoost.sec() << " seconds " << tvIncBoost.msec() << " mseconds" << std::endl ; std::cout << std::endl << usecsBoost << " micro seconds " << std::endl ; ACE_OS::system("pause"); return 0; } //////////////////////////////////////////////////////////////////////////////////////////////////////////// void gen_random_string(std::string &s, const int len) { static const char alphanum[] = "0123456789" "ABCDEFGHIJKLMNOPQRSTUVWXYZ" "abcdefghijklmnopqrstuvwxyz"; s.clear(); for (int i = 0; i < len ; ++i) { s.push_back(rand() % (sizeof(alphanum) - 1) ); } } Can anyone of you help me with this? Any help would be truly appreciated. Thanks in advance.

4 4

[task][fiber] question about memory leaks
by Dorogan' Konstantin 11 Nov '10

11 Nov '10

Sorry for my "good" english. :) I'am try to use boost.task 0.4.2 in my project. I found some leaks in the files handle.hpp and task.hpp related use_count member initialization. So boost::intrusive_ptr can't release allocated resources. After fixing promblem of the use_count initialization, i ran my unit tests again, but some leaks still living. I discovered that the problem is in the boost.fiber (v 0.5.1). If i try to run compiled unit tests of the boost.fiber manually, the garbage collector of the boost.test shows the leaks. In hopes of finding the place of data leaks, I perform these tests under Rational Purify, but Puify reports that no leaks were detected. Is the garbage collector of the boost.test lies, or the leaks is real!? Tested environment: Windows 7 x64, Windows XP boost 1.42.0 boost.context 0.2.0 boost.task 0.4.2 boost.fiber 0.5.1

1 0

[1.45] Release v trunk, as of r66483
by Jim Bell 11 Nov '10

11 Nov '10

Here's a list of all differences between release and trunk, as of r66483 (this morning). I won't slam the mailing list with its contents, so it's here instead: http://jc-bell.com/files/release-v-trunk-r66483.csv Organized by Folder (2nd column), and it might have a stray false positive or negative since I apply a patch to test. FYI.

5 10

Re: [boost] [msm] Event deferring causes stack overflow
by Christophe Henry 10 Nov '10

10 Nov '10

<snip> >It is not that nice for 2 reasons: > >1. It makes state machine behavior description more splinted in our case to several files ( not a big deal). >2. I have created our own defer functor which logs the event deferral with our logging framework >is there a way do similar function in a state like the no_transition for deferred events where I can implement some generic logging ?. Hi Richard, No, it's not nice. Luckily, last night brought fresh ideas and I implemented a fix. It has to run all regression tests to be 100% sure but it looks good so you can give it a try. For your logging functor, you have to provide a typedef inside the functor definition (like Defer is now doing in functor_row.hpp): typedef int deferring_action; This activates the normal deferred handling and the stack overflow is gone. The fix can be downloaded from the trunk (revision 66490). Unfortunately, it is too late to make it into the 1.45 so be careful if you switch to the new boost release :( I also didn't have time to actualize the doc yet. HTH, Christophe

1 0

[boost.log] Using stream manipulators to influence record formatting per sink
by Jamie Allsop 10 Nov '10

10 Nov '10

So I've been using the previously reviewed log library by Andrey Semashev with some degree of success. One thing however has caused me a great deal of trouble and I'm sure with a couple of changes would become a lot easier. Basically I have an application that outputs tabular data in the logs. I have objects that know how to stream themselves either as an ascii table or as an html table. I have a custom defined manipulator that can be set for a stream that the objects can check the value of, and stream themselves either as text or html as appropriate. In using the log library I face two major challenges. The first is how can I set the manipulator to the desired value so that the log record will be aware of the manipulator while building its message? It is not possible to do this by calling set_formatter() on the sink since the this receives *an already composed* record.message() string. My work-around for this is a lot more complex than I'd like it to be and I'll detail it here in the hope that it provides some insight to Andrey so that he might be able to offer a nicer alternative more in-keeping with his library. So starting at the beginning I have objects that implement streaming something like this: ---- template< class CharT, class Traits> friend std::basic_ostream< CharT, Traits>& operator<<( std::basic_ostream<CharT, Traits>& Ostream, const my_class& Value ) { if( log::output_format::value_in( Ostream ) == log::output_format::text ) { write_text( Ostream, Value ); } else if( log::output_format::value_in( Ostream ) == log::output_format::html ) { write_html( Ostream, Value ); } return Ostream; } ---- where 'value_in()' simply retrieves the value from the stream. Now during logging a temporary stream object is created that must have this manipulator applied. What I'd like to do is be able to specify if I want this manipulator applied on a per-sink basis. In much the same way as I can define a specific formatter for a sink, or a specific filtering rule for a sink. Currently there is no way to do that (at least not without a *lot* of effort) so I'll address this later. Instead I'll consider how we can deal with this be setting this desire globally by using a global attribute. We could write something like this: ---- log::output_format::type OutputFormat = log::output_format::html; std::pair< boost::log::core::attribute_set_type::iterator, bool > OutputFormatAttribute = boost::log::core::get()->add_global_attribute ( "OutputFormat", boost::make_shared< boost::log::attributes::constant<log::output_format::type> >( OutputFormat ) ); ---- Later we can extract this value to decide if we need the manipulator applied to the stream that is used to compose the log record message. We can use a helper object that is aware of all the manipulators that are to be applied and have it the log record stream based on attributes that represent them. For example, given a class called 'manipulators' we can define this insertion operator as: ---- template< class CharT, class TraitsT > inline std::basic_ostream<CharT,TraitsT>& operator<< ( std::basic_ostream<CharT,TraitsT>& Ostream, const manipulators& Manipulators ) { Manipulators.apply( Ostream ); return Ostream; } ---- where apply could be implemented (in our case) as this: --- template< class CharT, class TraitsT > void apply( std::basic_ostream<CharT,TraitsT>& Ostream ) const { boost::log::basic_record<CharT> Record( RecordHandle_ ); apply_output_format( Ostream, Record ); } private: ---- template< class CharT, class TraitsT > void apply_output_format( std::basic_ostream<CharT,TraitsT>& Ostream, const boost::log::basic_record<CharT>& Record ) const { log::output_format::type OutputFormat = log::output_format::text; boost::log::extract<log::output_format::type> ( "OutputFormat", Record.attribute_values(), boost::lambda::var( OutputFormat ) = boost::lambda::_1 ); if( OutputFormat == log::output_format::text ) { Ostream << log::set_output_format( log::output_format::text ); } else if( OutputFormat == log::output_format::html ) { Ostream << log::set_output_format( log::output_format::html ); } } ---- This then requires that the insertion operator is called as part of the streaming for any log record. In order to achieve this I need to write my own logging macros that do essentially what the severity logging macros do already but with the addition of streaming a 'manipulator' object at the same time so that the manipulators I want added can be applied to the stream that is constructed during that process. Here is the key macro that I need to write which gets substituted in a chanin of macros: ---- #define MY_LOG_STREAM_WITH_PARAMS_INTERNAL(logger, rec_var, params_seq)\ for (::boost::log::record_handle rec_var = (logger).open_record((BOOST_PP_SEQ_ENUM(params_seq))).handle(); !!rec_var; rec_var.reset())\ ::boost::log::aux::make_pump_stream((logger), rec_var).stream() << log::manipulators(rec_var) ---- The important part is this: ---- log::manipulators(rec_var) ---- The stream object is created and immediately passed an instance log::manipulators passing the RecordHandle to the constructor. Finally then I can have manipulators passed down to the stream that is used to create the log record message, and in my case allow the logged objects to either stream themselves as text or as html. A downside of this is that this only works with global attributes. So either all log sinks would one format or the other. Given that it is generally desirable to treat different sinks differently (and often the reason they exist) it would be much nice if there was the concept of sink-level attributes. These could be added to the record's attribute set along with the source and global attributes. Currently there is no way (it seems) to associate any state with a sink or make any decision during logging that relates to a sink. We can set a custom formatter that applies to the log message created but that has no effect on the record message itself. (Decorators only seem to be for post-processing of the record message). We can also set a custom filter to determine whether a log record gets logged by a given sink. It seems odd then that we have no way to to be able to say (for example) that a given sink will have an attribute that indicates that all messages through that sink should be in a certain format (for those objects that support the formatting). From looking at the code this does not appear to be a major issue so hopefully it could be supported. It would also be highly desirable to be able to specify manipulators directly that could then be applied to the record message stream itself. Supporting this at the global and sink level should be possible. If we had sink-level attributes the technique I described could be used to simulate this, even though it is somewhat kludgey. Most directed to Andrey, but I'd welcome comments and suggestions from others familiar with the log library. Jamie

2 3

[icl] The Interval Container Library (ICL) moved to the Boost trunk
by Joachim Faulhaber 10 Nov '10

10 Nov '10

Dear All, thank you again for you welcoming comments on the latest version of the now renamed Interval Container Library ICL (formerly Interval Template Library ITL). And thanks to Hartmut for your "invitation" to move the library to the Boost trunk. This might be a small step for Boost but it's a big step for me ;-) I am not sure what has to be done now, so my test suite will be executed in the daily tests that is summarized in the trunk summary on the Boost web-site. I have added my lib into the list in status/Jamfile.v2. I am also wondering how I get information about and deal with errors from those tests on platforms that I can not run on my machine. I didn't find anything about those topics on the Wiki so far. Maybe someone can provide some guidance :) Thanks, Joachim

1 0