November 2007 - Boost-users - lists.preview.boost.org

ublas::vector<A>
by e r 07 Nov '07

07 Nov '07

hi, is it possible to do ublas::vector<A> vec; where A is a class that i define, not a built in type? if so, what are the requirements on A?

1 0

boost.python switching from v2.5 to v2.4
by desai3＠cooper.edu 06 Nov '07

06 Nov '07

Hi all: I have used Boost.Python successfully to wrap around C++ API in python2.5. I must say that boost.python is an amazing library. However, if I want to create the same API for python2.4 do I have to rebuild the boost with python2.4 and then run bjam again, or is there any simple solution than that? The reason for asking this is because changing the Jamrules file did not help much. Thank you for your help. Maulik

2 1

[filesystem] Apparent non-compliant code in current operations.hpp, ...
by Marty Fried 06 Nov '07

06 Nov '07

While building 1_34_1 on various platforms we support, I came across some compile errors for the AIX compiler, xlC/vacpp that weren't caught by gcc or MSVC, possibly due to being hidden by a level of indirection. The errors were mostly undefined classes where the template class was fully defined at the time of instantiation. I apologize if my explanation isn't clear or totally correct - I'm not an expert on the C++ standard, and got a lot of the actual explanation from coworkers that know it better than I do. But it's up to me to do the actual reporting and to try to get it to work for us on all the platforms we support. My original post was a few days ago, with the subject "[filesystem] vacpp errors building 1_34_1 filesystem library on AIX". I am because it may look like a vacpp problem, which is an unsupported compiler, but it's not really. It just happens that xlC was the only compiler that caught the error, I think. My original post can also be read at this link: http://permalink.gmane.org/gmane.comp.lib.boost.user/31028 The problem is that the class basic_directory_iterator is not fully defined until after it is used by remove_all(). There were a few other similar errors which I can provide if someone takes an interest in this. If I am wrong, or if there is a reason this won't be fixed, it would be helpful to know, so we can decide what to do. -- Marty Fried Senior Software Developer, Aldon USA http://www.aldon.com

3 7

boost timers
by Mark Wyszomierski 06 Nov '07

06 Nov '07

Hi, Do any of the boost timer methods allow elapsed time to be measured in a separate thread? Something like: int main() { boost_thread timer; timer.Start(); while () { // some timeout. sleep(5000); // check how much time has passed. cout << timer.GetElapsedTime() << endl; } return 0; } Thanks

2 2

[program options] storing values into a set instead of vector?
by Ovanes Markarian 06 Nov '07

06 Nov '07

Hello all! I would like to allow only unique parameters to be stored: param = X param = Y param = X should result in a collection of {X, Y} Therefore I want my params to be stored in a set instead of vector? Is there any chance to implement/specify it without putting the params into the vector of strings and re-inserting them into a set? Doing it the way: value<string_set>() result in an error. Some other question? As I see "program options" works only with basic_strings<charT>, where it is assumed that the traits class is default. What if I wish to specify my traits class for the basic_string will it work? Many thanks! Ovanes

1 0

Starting ThreadClass including Mutexes
by Honki 06 Nov '07

06 Nov '07

Hi boost-users, I'm new in the mailing list, so I hope, that I'm right here and you can help me. I want to start a ThreadClass that includes some boost::mutex variables, but it doesn't work because of the "noncopyable" definition of the mutex-class. *******************MyClass******************* ...... class OutputProcessor_C { public: ClientOutputProcessor_C(); virtual ~ClientOutputProcessor_C(); void operator()() { while( true ) { boost::xtime wakeUpTime; boost::xtime_get(&wakeUpTime, boost::TIME_UTC); wakeUpTime.sec += ( 0.05 ); boost::thread::sleep( wakeUpTime ); Process(); } } protected: void Process(); boost::mutex m_publishDataMutex; ...... ********************************************* *******************MyMain******************** ...... int main(int argc, char* argv[]) { // Classpointer boost::shared_ptr<OutputProcessor_C()> outputProcessor(new OutputProcessor_C()); // generate Thread boost::shared_ptr<boost::thread> threadOutputProcessor(new boost::thread(*outputProcessor)); ...... ********************************************* When I tried to compile the source code, there is an error - "It is not possible to copy the constructor for the class 'boost::mutex'". Where is the problem? I don't want to copy anything. If it is not possible to generate a simple thread class that includes some mutexes, that is really weak. Greetings, Matthias

2 1

[thread] static mutex construction
by Pedro Lamarão 06 Nov '07

06 Nov '07

Hello all, I've been wondering how boost::mutex plays in the static initialization game. I'm toying with a "plugin" system where plugins statically register with a singleton plugin_manager using a helper struct like this: template <typename PluginT> struct register_plugin { register_plugin () { shared_ptr<plugin_base> ptr(new PluginT); plugin_manager::instance().register(ptr); } }; static register_plugin objects, specialized for each plugin type, are constructed in an anonymous namespace to register the plugin at static initialization time, like so: namespace { register_plugin<foo_plugin> rp; } plugin_manager::instance() looks like: { static plugin_manager* s_instance; // s_mutex is a static member { boost::mutex::scoped_lock lock(s_mutex); if (s_instance == 0) { s_instance = new plugin_manager; } } return *s_instance; } So far I've implemented two plugins in this application and everything just works. It seems to me there is a static initialization race between every register_plugin object and plugin_manager::s_mutex. Am I just lucky or is there some magic about mutexes in this regard? I'm working on a Fedora 7 i386 system. -- Pedro Lamarão

2 1

[crc] Easy way to get the "correct checksum" constant?
by Norbert Unterberg 06 Nov '07

06 Nov '07

Some CRC algorithms (when using 0xFFFF as an initial value and final XOR) have a nice property that then checking the consistency of a data block that has a CRC checksum appended, the resulting checksum value is a constant. Is there an easy way to extract this constant value from boost::crc directly? Example: We use a CRC like this: typedef boost::crc_optimal<16, 0x1021, 0xFFFF, 0xFFFF, true, true> crc_type; When sending a block of data, we append the generated CRC. The receiver side feeds the complete block incluing the CRC bytes into its generator. When the data block has no errors, the crc.checksum() is always exactly 0x0F47. Currently, I manually calculate that constant by something like this: crc_type crc; crc_type::value_type val(0); crc.process_bytes(&val, sizeof(val)); m_key = crc.checksum(); Is there an easier way to extract this "crc is correct" value? Norbert

1 0

boost-users [boost::function] template.
by chun ping wang 04 Nov '07

04 Nov '07

Hi I have a problem with boost::function in association with template. I have a gain function declaration declare as this.. template < template <typename ELEM,typename = std::allocator<ELEM> > class CONT, typename T, typename U > long double gain(const CONT<boost::unordered_map<T, U> >& data, const T& attr, const T& target_attr); // old signature does compile template < template <typename ELEM,typename = std::allocator<ELEM> > class CONT, typename T, typename U > boost::any ID3( const CONT<boost::unordered_map<T, U> >&, const T&, const CONT<T> &, boost::function<long double (const CONT<boost::unordered_map<T, U> >&, const T&, const T&) > ); // new signature DOES NOT compile. template < template <typename ELEM,typename = std::allocator<ELEM> > class CONT, typename T, typename U > boost::any ID3( const CONT<boost::unordered_map<T, U> >&, const T&, CONT<T> *, boost::function<long double (const CONT<boost::unordered_map<T, U> >&, const T&, const T&) > ); In the main file i have things pass as this. // old way of passing works with old signature boost::any tree( MyAnn::ID3<std::vector, std::string, std::string>(data, target_attr, attribute, MyAnn::gain<std::vector, std::string, std::string>) ); // line 76 to 82. boost::function<long double (const std::vector<boost::unordered_map<std::string, std::string>&, const std::string&, const std::string&)> gainfunc; gainfunc = MyAnn::gain<std::vector, std::string, std::string>; boost::any tree( MyAnn::ID3<std::vector, std::string, std::string>(data, target_attr, boost::addressof(attribute), gainfunc) ); I compile the file and get teh following errors. Can anyone tell me why changing the orginial signature from const CONT<T>& to CONT<T> * all of sudden breaks the code. Thanks.

1 1

回复： [Boost-users] "Getting Started" document : missing documentation of the stdlib flag
by 鑫赖 04 Nov '07

04 Nov '07

Hi: Thank you.I try it! _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users ___________________________________________________________ 雅虎邮箱，终生伙伴！ http://mail.yahoo.com.cn/

1 0