September 2007 - Boost-users

[multi-index] possible extension
by Ovanes Markarian 21 Sep '07

21 Sep '07

Hello Joaquín, I would like to make a proposal for additional feature of optionaly making multi-index container non-copyable (via copying policy class). The same way as a user can decide which indicies to use she/he could specify whether the new data-structure should be copyable or not. In my case I was assuming to pass container by reference, but it was copied by some function object. Which produced despite the performane issue wrong results. Using the non-copyable policy I would simply get a compiler error. With Kind Regards, Ovanes

2 1

[python]
by Gennadiy Rozental 21 Sep '07

21 Sep '07

Hi, In my new company we are using Boost.Python. Boost version is 1.33.1. It's my first steps with the library, so please bare with me if this is something trivial. I've got this noncopyable class Foo class Foo : boost::noncopyable { /*...*/ }; and shift operator defined: Foo& operator<<( Foo& foo, int ) { ... return foo; } Here is my export statement (FooSP is smart pointer on Foo using namespace boost::python; class_<Foo, FooSP, bases<FooBase>, boost::noncopyable>( "Foo", no_init ) .def( self << int() ) ; Here is my Python script extract that is being executed from C++ "foo << 2 << 4" I never got to the second shift call. Instead I an getting exception from Boost.Python: TypeError: No to_python (by-value) converter found for C++ type: class A I kinda found couple ways to deal with it: 1. define operator<< to return FooSP, which I do not like for assymetry and other reasons 2. avoid using self << int() and define explicit function: Foo& shift_op( Foo& foo, int ) { ... return foo; } ... .def( "__lshift__", &shift_op, return_internal_reference<1>() ) What is the proper way to deal with it? And why original error is reported? Gennadiy

2 1

[mpl] Instantiating a vector of factory functions
by Dan Posluns 21 Sep '07

21 Sep '07

Okay, I'm a lot closer to my goal of having an automated factory of objects that are dynamically polymorphic but have their inheritance set up through static means. What I have now is something like this: // My factory template template<class Base, class Derived> struct Factory { Base *create() { return new Derived(); } }; // My vector of factories for the different Derived types typedef mpl::vector< Factory<Base, Derived1>, Factory<Base, Derived2>, Factory<Base, Derived3> > MyFactories; (I actually use MPL to automatically cross and combine different classes, which is pretty cool.) Now I could create a vector of the same length with every type as a pointer to a Factory function, and with Fusion I could probably instantiate it (although I haven't got that working yet). But that still wouldn't get me what I want, which is a table of those factory functions that I can index into at run-time. Can anyone suggest a technique for doing this? Thanks, Dan. -- Dan Posluns, B. Eng. & Scty. (Software Engineering and Society) dan(a)danposluns.com - ICQ: 35758902 http://www.danposluns.com "The great thing about being the only species on the planet that makes a distinction between right and wrong is that we get to make the rules up for ourselves as we go." - Douglas Adams, Last Chance to See

2 1

[date_time/general] license question
by Olaf Krzikalla 21 Sep '07

21 Sep '07

Hi all, In the date_time library there is the libs\date_time\xmldoc\license.xml covering the date_time library. It states that the copyright notice of the date_time library has to "appear in supporting documentation". Does this includes manuals for commercial binaries? I m'wondering because IMHO this wouldn't comply with point 4 of the boost license requirments as written at www.boost.org/more/license_info.html#requirements OTOH I'm neither a lawyer nor a native english speaker so I'm sure I misinterpret something. Can anyone shed some light on this? Thanks Olaf Krzikalla

2 1

Some errors in Boost.Log
by Jens Seidel 21 Sep '07

21 Sep '07

Hi, I know that Boost.Log is not an official Boost library but I hope that this is the right place to collect infos and errors of this library (especially as the author does not respond). I always missed the fact Boost.Log is not able to redirect messages of a special priority (say err) to a file. Redirection is only based on the logger name (e.g. "project.file_accesses"). It is possible to match all priorities higher than a choosen one but if you collect many debug outputs you will miss a few errors in the large log. Now I noticed that I can just create separate loggers such as logger_err and output an err message to the specified logger and additionally to logger_err. Now I'm able to redirect err-level outputs via logger_err *only*, great! I missed a solution like this in the library (yep, it's not difficult, but ...) and it may be useful for others too. Nevertheless the documentation manipulating_logs.html or the library contains errors: <quote> Here are the predefined levels: info (1000) warn (1200) dbg (1400) err (1600) fatal (2000) disable_all ( std::numeric_limits<unsigned>::max() ) default_level (1000) enable_all (0) </quote> First of all dbg should have a much lower priority than info, not a higher one! Also it should be disable_all (0) enable_all ( std::numeric_limits<unsigned>::max() ) In the past I noticed (and reported to the author who ignored my mail) also trivial compile errors with latest g++ compiler (missing ";", ...). I will continue reporting issues to this list, just in case another person wants to maintain Boost.Log or the author will continue his work (which I hope) ... What logger framework do Boost users use? I do not remember reading much about Boost.Log on this list ... Jens

1 0

Re: [Boost-users] Getting errors while compiling a project
by Vitali Gontsharuk 21 Sep '07

21 Sep '07

It seems that the moderator didn't want a message that long to be published, so I had to delete the full log from the message, but it is still available from the first message in this thread. You just have to download it and open it from the disk. Ok, i saw that the log file is not viewable directly (wrong MIME-type > specified by server). > So i copied it as text to the end of this message. > > The uppermost error appears in the the "fusion/support/detail/access.hpp" > file in the 23. line, here a couple of the relevant code lines: > > namespace boost { namespace fusion { namespace detail > { > template <typename T> > struct ref_result > { > typedef typename add_reference<typename T::type>::type type; // > <----- THE ERROR IS THROWN HERE > }; > ... > } > > Here the correspoding error messages from the log: > 1>D:\boost\include\boost/fusion/support/detail/access.hpp(23) : error > C2143: syntax error : missing ';' before '<' > 1> D:\boost\include\boost/fusion/support/detail/access.hpp(24) : > see reference to class template instantiation > 'fus_v2::boost::fusion::detail::ref_result<T>' being compiled > 1>D:\boost\include\boost/fusion/support/detail/access.hpp(23) : error > C4430: missing type specifier - int assumed. Note: C++ does not support > default-int > 1>D:\boost\include\boost/fusion/support/detail/access.hpp(23) : error > C2039: 'type' : is not a member of '`global namespace'' > 1>D:\boost\include\boost/fusion/support/detail/access.hpp(23) : error > C2238: unexpected token(s) preceding ';' > > > But originally, the error comes from the following definition in our > "cmd_processor.hpp" file: > /// anonymous namespace for internal linking > namespace > { > /// iterator through message definition files > /// this class is designed to iterate sequentially through > directories containing message definition files > /// and through single file specifications additionaly defined, so it > implements the iterator pattern and treats > /// directory content and additional single file specification as one > iterable sequence > struct msgdef_file_iterator > : boost::iterator_facade > < > msgdef_file_iterator > , fs::path > , boost::forward_traversal_tag > > > { > ... > private: > /// type definition for a variant containing either nothing > (empty_type) or directory_range or file_range > typedef boost::variant<empty_type, directory_range, file_range> > range_holder; // <----- range_holder DEFINITION IS HERE > > ... > private: > /// configuration settings > cmdproc::config const* conf_; > /// current range holder > range_holder holder_; // <----- THE ERROR > IS THROWN HERE > /// current path which is used from the range > mutable fs::path curr_path_; > }; > ... > } > > Here the corresponding error messages: > 1> D:\boost\include\boost/variant/variant.hpp(1055) : see reference > to class template instantiation > 'boost::detail::variant::find_fallback_type<Types>' being compiled > 1> with > 1> [ > 1> > Types=boost::mpl::l_item<boost::mpl::long_<3>,fus_v2::boost::fusion::pair<`anonymous-namespace'::msgdef_file_iterator::tag_empty,`anonymous-namespace'::msgdef_file_iterator::tag_empty>,boost::mpl::l_item<boost::mpl::long_<2>,fus_v2::boost::fusion::pair<`anonymous-namespace'::msgdef_file_iterator::tag_directory,`anonymous-namespace'::msgdef_file_iterator::msgdef_dirs>,boost::mpl::l_item<boost::mpl::long_<1>,fus_v2::boost::fusion::pair<`anonymous-namespace'::msgdef_file_iterator::tag_file,`anonymous-namespace'::msgdef_file_iterator::string_iterator_range>,boost::mpl::l_end>>> > > 1> ] > 1> ..\..\..\conf\cmd_processor.cpp(400) : see reference to class > template instantiation 'boost::variant<T0_,T1,T2>' being compiled > 1> with > 1> [ > 1> T0_=`anonymous-namespace'::msgdef_file_iterator::empty_type, > 1> > T1=`anonymous-namespace'::msgdef_file_iterator::directory_range, > 1> T2=`anonymous-namespace'::msgdef_file_iterator::file_range > 1> ] > > I hope, this can help you to be able to help me :) > I really need your help!!! > If you need any further information, please, let me know! > > Thanks in advance! > > Vitali Gontsharuk >

1 0

Upgrading to 1.34.1
by Hughes, James 21 Sep '07

21 Sep '07

Hi All, I am having trouble getting bjam to run on my Ubuntu 6.06 system - its complaining about a missing GLIBC_2.4. I'm downloading this from gnu (fingers crossed), but wondered if there is a quick 'Ubuntu' way of installing the latest bjam and associated glibc dependencies that anyone is aware of? James This message (including any attachments) contains confidential and/or proprietary information intended only for the addressee. Any unauthorized disclosure, copying, distribution or reliance on the contents of this information is strictly prohibited and may constitute a violation of law. If you are not the intended recipient, please notify the sender immediately by responding to this e-mail, and delete the message from your system. If you have any questions about this e-mail please notify the sender immediately.

2 2

Re: [Boost-users] Using multi-index as container type in templated class
by Joaquín Mª López Muñoz 21 Sep '07

21 Sep '07

Hi Paul, Paul.VanHagen(a)shell.com ha escrito: > Hello, > > I am having trouble using a boost::multi_index_container type in a template class where the > multi-index value type is depending on the template arg type of the class. > > In particular, I'm getting compiler errors when trying to define the iterator and index types > using the "type" typedefs in the multi_index container. Apparently, at compile time the > index and iterator types cannot be deduced resulting in errors like "cannot convert to > typename, etc.". > > I'm using GCC 3.3.1 and Boost 1.34.1. Well, seems like the compiler is having a hard time with dependent nested types. Probably newer versions of GCC will pose less problems. Fortunately, you can use global variants for nth_index etc., which happen to be easier for the compiler. So, instead of typedef typename mi::nth_index<0>::type mi_by_value; you can write typedef typename nth_index<mi,0>::type mi_by_value; Besides, you write in several places: m_c.get<...>() but as the calls are in dependent contexts you have to write them like this instead: m_c.template get<...>() or like this, as you're alreading doing in MultiIndex::MultiIndex(): get<...>(m_c) The attached code includes all the corrections above and have been verified to compile and run OK with GCC 3.2. Hope this helps, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

2 2

[boost-users] Library names with mingw-g++ onm windows
by John Femiani 21 Sep '07

21 Sep '07

How does one invoke bjam on the version of boost in SVN-Trunk so that mingw-g++ libraries use '.a' as a suffix instead of '.lib' ? When I use the mingw g++ compiler on Windows I expect it to produce libraries ending with '.a' since that is what the compiler typically looks for when linking with something like: -lboost_filesystem-mt-d Granted I could just put the whole library name -llibboost_filesystem_mt-d.lib but is there a way to make bjam produce .a as a suffix instead? Shouldn't this be the default behavior if I am using a gcc based toolset? -- John

2 1

Re: [Boost-users] [multi-index]/shared_ptr template issue -compiles on MSVC8 but not gcc (solved)
by Mark Schlegel 21 Sep '07

21 Sep '07

>> // this AGE_INDEX does not on gcc but DOES compile on MSVC8 >> typedef AgedMap<T>::AGEDMAPTYPE::index<agetag>::type AGE_INDEX; >The correct syntax is >typedef typename AgedMap<T>::AGEDMAPTYPE::template > index<agetag>::type AGE_INDEX; >(note the intervening template keyword) which, just for >succintness, in your particular ought to be the same, I think, as > ..... The above works (combo change of typename and ::template) so I've settled on using for brevity: typedef typename MIContainer::template index<agetag>::type AGE_INDEX; typedef typename AGE_INDEX::iterator AGE_ITER; thanks, Mark

1 0