August 2006 - Boost-users - lists.boost.org

Jamfile for app using date_time
by Steve Hyatt 29 Aug '06

29 Aug '06

Hi all. I am trying to write a Jamfile for a simple test app which uses the boost date_time library. I am unable to figure out how to tell the boost build system to link in that library. I have a Makefile which can do the link: $make g++ -c -pipe -Wall -Wno-unknown-pragmas -DLINUX -I./ -g -o tester.o tester.cpp g++ -o tester tester.o -lboost_date_time-gcc-1_33_1 Can someone show me a Jamfile that 'just works' (TM). Thanks Steve saphlyx(a)shaw.ca

2 1

Interprocess bug and patch
by Adam Lewis 28 Aug '06

28 Aug '06

Not sure if this is the right place for this bug but I believe it is. The function get_max_num_msg() was returning the incorrect member.

2 1

asio and interprocess fifo
by Koen Deforche 28 Aug '06

28 Aug '06

Hello, For a server-client application, I am experimenting with boost::asio and boost::interprocess -- this is my first intensively boostified project. Since boost::interprocess is work-in-progress, I am using the CVS version. I have found some smaller problems (see at the end of my email), but so far, all works well. I have noticed that in interprocess, an ipc/named_fifo is planned. I was wondering if there will be plans for the asio library to create an asynchronous I/O service for such a fifo ? If not, is this feasible ? I did not find any documentation on how to write a new io_service in the asio documentation ? Regards, koen 1) Timeouts always timed out immediately, which was caused by : interprocess/sync/ptime_to_timespec.hpp line 29: ts.tv_sec = duration.seconds(); should be ts.tv_sec = duration.total_seconds(); 2) Compile problem in interprocess/errors.hpp, which was solved by : interprocess/errors.hpp line 86: void fill_system_message( int system_error, std::string &str) should be inline void fill_system_message( int system_error, std::string &str) (probably the same is true for the BOOST_WINDOWS versino on line 65 ?

2 1

[statechart] No state?
by Johan Nilsson 28 Aug '06

28 Aug '06

Hi, I'm implementing an asynchronous state machine and have problems with events not being dispatched. All event thrown at the fsm are dispatched to the unhandled_event method. When debugging the problem, I used typeid on the event to see that the expected events were generated (they were) from the event dispatch threads. After this, I tried to get the current state from within the unhandled_event() method - but to my surprise the statement "this->state_begin() != this->state_end()" is never satisfied. Looks like there's no current state. What's even more confusing is that this behaviour is only exposed in the application itself. This part of the project is divided into three modules; the application, a static library containing most of the code, and a unit test driver. The test driver and the application links to the library. When running the unit tests, the dispatching works and it is possible to see the current state using the above method. Release or debug mode doesn't matter. RTTI is enabled. Using RC_1_34_0 and VC8. What could possibly expose this behaviour? // Johan

2 4

How to wrap a template function(not a member of a template class) into python?
by Qinfeng(Javen) Shi 28 Aug '06

28 Aug '06

Dear All, How can I wrap a c++ template function(not a member of any class) by pyste or manually? /**************************************/ for example : template<class T> void push(v_array<T>& v, const T &new_ele) { while(v.index >= v.length) { v.length = 2*v.length + 3; v.elements = (T *)realloc(v.elements,sizeof(T) * v.length); } v[v.index++] = new_ele; } /***************************************/ Here v_array is a class. I know how to wrap a c++ template class into python by pyste. But what should I do if the template function is not a member of a class. (I don't want to put the function"push()" into a class). Any suggestion will be appreciated. Thanks! Javen Research School of Information Sciences and Engineering Australian National University Locked Bag 8001 Canberra ACT 2601

2 1

Re: [Boost-users] boost program_options
by Vladimir Prus 28 Aug '06

28 Aug '06

On Monday 28 August 2006 15:49, you wrote: > Hello, Hi, please send all Boost questions to some mailing list, not in private. > > in your tutorial for program_options at the boost site you talk about > the method allow_unregistered() lite this: > > parsed_options parsed = > command_line_parser(argv, > argc).options(desc).allow_unregistered().run(); > > > However, I'm using boost 1.33 and the command_line_parser class uses > private inheritence from the detail::cmdline class, which makes the > allow_unregistered() method unavailable for public use. What is the > deal? That's a bug in 1.33. The operation of allow_unregistered was unit-tested, but since unit test used detail::cmdline directly, the fact that the method is not accessible in command_line_parser went undetected. I believe it's fixed in CVS. - Volodya

2 1

[interval]
by Andy Little 28 Aug '06

28 Aug '06

Shouldnt this work?: #include <boost/numeric/interval.hpp> int main() { boost::numeric::interval<double> v(.5); cos(v); } ------------------ I get this error in VC7.1: c:\boost\include\boost-1_33_1\boost\numeric\interval\transc.hpp(73) : error C2039: 'cos_down' : is not a member of 'boost::numeric::interval_lib::rounded_math<double>' c:\boost\include\boost-1_33_1\boost\numeric\interval\hw_rounding.hpp(53) : see declaration of 'boost::numeric::interval_lib::rounded_math<double>' e:\Projects\Test\test.cpp(6) : see reference to function template instantiation 'boost::numeric::interval<T> boost::numeric::cos<double,boost::numeric::interval_lib::default_policies<T>::type>(const boost::numeric::interval<T> &)' being compiled with [ T=double ] c:\boost\include\boost-1_33_1\boost\numeric\interval\transc.hpp(73) : error C2039: 'cos_up' : is not a member of 'boost::numeric::interval_lib::rounded_math<double>' c:\boost\include\boost-1_33_1\boost\numeric\interval\hw_rounding.hpp(53) : see declaration of 'boost::numeric::interval_lib::rounded_math<double>' c:\boost\include\boost-1_33_1\boost\numeric\interval\transc.hpp(75) : error C2039: 'cos_up' : is not a member of 'boost::numeric::interval_lib::rounded_math<double>' c:\boost\include\boost-1_33_1\boost\numeric\interval\hw_rounding.hpp(53) : see declaration of 'boost::numeric::interval_lib::rounded_math<double>' regards Andy Little

2 3

boost::condition
by Andrew Wingorodov 28 Aug '06

28 Aug '06

I wish to lock some threads up to the end of one of them, and after to start all of them without restrictions. show to me an example, plz the timedwait is necessary. -- andr.ru

1 0

[serialization] class versioning with template classes
by Oliver Mutz 28 Aug '06

28 Aug '06

Hi all, has anyone ever used the versioning mechanism of the serialization library together with template classes? What I want to do is something like: class TemplateA { public: TemplateA(); public: ... //Serialization stuff private: friend class boost::serialization::access; template<class Archive> void save(Archive & ar, const unsigned int /*version*/) const { ar & boost::serialization::make_nvp("HighValue",m_high) & boost::serialization::make_nvp("NewData",m_newData); } template<class Archive> void load(Archive & ar, const unsigned int version) { ar & boost::serialization::make_nvp("HighValue",m_high); if (version > 0) { ar & boost::serialization::make_nvp("NewData",m_newData); } } BOOST_SERIALIZATION_SPLIT_MEMBER(); private: T m_high; T m_newData; }; // class TemplateA BOOST_CLASS_VERSION(TemplateA, 1); The above won't compile for sure, but I'm looking for a way to use the versioning without having to specialize it for each type T which I'm going to use this class with. Best regards Oliver Mutz

1 0

Debian Sarge less-then-g++-4.0.x link-time issues
by Matt England 28 Aug '06

28 Aug '06

Our project is experiencing a lot of Debian link-time issues with Boost libraries, namely with g++ versions less then 4.0.x. Some bug listings: http://www.cleversafe.org/bugs/show_bug.cgi?id=260 http://www.cleversafe.org/bugs/show_bug.cgi?id=239 Does the boost-users community have any pointers/ideas/suggestions/references? Some of our devs think that these problems may be due to bugs in g++ 3.3.5 or less; we have unconfirmed reports that g++ 3.3.5 or less builds may have failed on other Linux platforms. -Matt Dispersed Storage: http://cleversafe.org

1 1