January 2009 - Boost-users - lists.preview.boost.org

Using iostream to create ZIP archive
by Peter Json 14 Jan '09

14 Jan '09

Hi, I'm trying to use boost to create a "zip" file containing more than one file. To my understanding I should use boost::iosteams to do this. In the boost documentation I there is only an example on how to unzip a file ..... http://www.boost.org/doc/libs/1_37_0/libs/iostreams/doc/classes/zlib.html#e… ... but I dont understand how I could do the opposite (create a zip-archive) Especially how to do the naming of the files inside the archive. Does anyone have an example on howto do create a zip-archive containing a couple of text files? Thanx /Peter/ -- View this message in context: http://www.nabble.com/Using-iostream-to-create-ZIP-archive-tp21461391p21461… Sent from the Boost - Users mailing list archive at Nabble.com.

2 1

large enum and native array indexing
by Hicham Mouline 14 Jan '09

14 Jan '09

hello, we have a 3rd party lib which provides an enum OP with some hundred members, but each member is defined with numbers that are not contiguous. enum OP { op1=56, op2=6, op3=5 ..... }; We have a performance critical function f that looks like this: status f( double& result, OP , int thing ); typical usage calls f( result0, op0 ) then f(result1, op1) .... f(resultn, opn) on the same thing. these calls, if merged, could have some speedup as there is some repetitive work done for say op3 and op7. I then thought of doing void f( std::map< Op, pair<status, result> >& output, const std::set<OP>&, int thing ) to collapse the above calls to 1. Inside this f, we set results in output with the [] operator. then I compared this to void f( double results[], status statues[], const std::set<OP>&, int thing ), thanks to the native array contiguous access, it was 4 times faster. therefore i wrote a numerical pseudo-metafunction "Position" to take the enum OP as an input and return a size_t, reflecting the index for the array. A call would be like: std::set<OP> opset; opset.set( op1 ); opset.set( op2 ); ... opset.set( op7 ); double results[7]; status statuses[7]; f( results, statuses, opset, thing ); // then to read output results[ Position<OP5>::value ]; Now, the question is about defining the metafunction Position. template< OP op > struct Position {}; template<> struct Position<OP5> { const size_t value=4; }; .... The 3rd party enum OP in the header file I can't change. Is there a way to generate my metafunction automatically? Any facility in mpl or fusion instead of defining my own metafunction? regards,

2 1

[ASIO/1.35.0]: Exception thrown in epoll_reactor::run (async API)
by Slawomir Lisznianski 14 Jan '09

14 Jan '09

Running on RedHat Linux, compiled with gcc 4.x. Program terminates with "Transport endpoint is not connected" -- io_service::run (processing loop) ends with exception thrown. I'm using asynchronous API. Stack trace of optimized debug (-02 -g) binary shows: #0 0x0000003a15a30155 in raise () from /lib64/libc.so.6 #1 0x0000003a15a31bf0 in abort () from /lib64/libc.so.6 #2 0x0000003a15a6a38b in __libc_message () from /lib64/libc.so.6 #3 0x0000003a15a71634 in _int_free () from /lib64/libc.so.6 #4 0x0000003a15a74c5c in free () from /lib64/libc.so.6 #5 0x00000000005093b0 in boost::asio::detail::epoll_reactor<false>::cleanup_operations_and_timers () #6 0x000000000050e6aa in boost::asio::detail::epoll_reactor<false>::run () #7 0x000000000050ebe3 in boost::asio::detail::task_io_service<boost::asio::detail::epoll_reactor<false> >::run () The problem occurs sporadically, under moderate load (about 30 client connections). Appears to be coming from cleanup_operations_and_timers() I can catch exceptions leaking from io_service::run, but would appreciate more info on why this occurs and whether it has other side-effects. Thanks, Slawomir

1 0

Boost graph add vertex
by Damian 14 Jan '09

14 Jan '09

Just as in add_edge(1,2,graph) you get to choose the numbers which represent the vertex is it possible to do the same with add_vertex where i actually choose the vertex number. I guess the syntax might be like add_vertex(3,graph) Cheers Damian

2 1

[boost-users][Milti-Index] labmda and composite key
by Igor R 14 Jan '09

14 Jan '09

Hello, I've got the following data structure: struct Service; typedef shared_ptr<Service> ServicePtr; struct Event { ServicePtr service() const; ptime time() const; }; typedef shared_ptr<Event> EventPtr; struct Cache { struct time_tag{}; typedef mi::multi_index_container< EventPtr, mi::indexed_by< mi::ordered_non_unique< mi::tag<time_tag>, mi::composite_key< EventPtr, mi::const_mem_fun<Event, ServicePtr, &Event::service> mi::const_mem_fun<Event, pt::ptime, &Event::time> > > > > type; typedef type::index<time_tag>::type ByTime; }; Cache cache_; ServicePtr servicePtr = ...; period = ...; I try to extract all the events belonging to some service and bounded by "period": namespace bl=boost::lambda; pair<Cache::type::iterator, Cache::type::iterator> byPeriod = cache_.get<Cache::time_tag>().range(bl::_1 >= make_tuple(servicePtr, period.begin()), bl::_1 <= make_tuple(servicePtr, period.last())); The cache contains lots of matching events, but for some reason, this expression always returs empty region, i.e. byPeriod.first==byPeriod.second. So I'd like to know if the above expression is correct or I'm doing something wrong? Thank you.

2 3

[gil] any_image* conversion to concrete type
by Jan Boehme 14 Jan '09

14 Jan '09

Hi, the following sample code introduces my problem: #include <boost/gil/extension/dynamic_image/dynamic_image_all.hpp> #include <boost/mpl/vector.hpp> #include <boost/gil/extension/io/jpeg_dynamic_io.hpp> int main() { using namespace boost::gil; typedef boost::mpl::vector<rgb8_image_t, cmyk16_planar_image_t> images_t; typedef any_image<images_t> any_image_t; typedef any_image_view<boost::mpl::vector2<rgb8_view_t, cmyk16_planar_view_t > > any_view_t; typedef any_image_view<boost::mpl::vector2<rgb8c_view_t, cmyk16c_planar_view_t> > any_const_view_t; any_image_t img; jpeg_read_image("test2.jpg", img); any_const_view_t cview( img ); rgb8c_view_t rgb_view = cview; } In the last line I want to convert the any_const_view to a concrete type what runs fail in lack of a member named 'pixels'. Does anyone know about that kind of conversion as the documentation imho doesn't contain such a sample. Right now I have to copy the pixels in the following form but I can't believe that this is necessary: rgb8_image_t rgb_image(img.dimensions()); copy_pixels(color_converted_view<rgb8_pixel_t>(const_view(img)), view(rgb_image)); rgb8c_view_t rgb8c_view( const_view(rgb_image) ); Thanks a lot, Jan.

1 0

multi_array resize question
by Wilfried Holzke 14 Jan '09

14 Jan '09

Hi, The documentation from multi_array says that if elements are added while reszize is called, the new elements are initialized with the default constructor. I have the following: typedef boost::multi_array<Object*, 2> Array_Type; What is the default constructor of a pointer? Or do need to do the following: class ptr { public: ptr() { p = 0; } Object *p; } typedef boost::multi_array<ptr, 2> Array_Type; ... to get all new elements to bee zero-pointers? Thanks for help. regards W. Holzke

3 2

[serializaton] namespace issues with extended type info
by Daniel Roberts 14 Jan '09

14 Jan '09

If I nest macro calls such as: namespace SomeNamespace { BOOST_SERIALIZATION_ASSUME_ABSTRACT(SomeClass) BOOST_CLASS_TYPE_INFO(SomeClass, extended_type_info_no_rtti<SomeClass>) BOOST_CLASS_EXPORT(SomeClass) } inside the namespace containing the type I get a TON of compiler errors. The wackiest is that the compiler thinks boost is a namespace inside of my namespace. Things improve (not fully tested) if I move the macros outside of my namespace and qualify the type with namespace name: BOOST_SERIALIZATION_ASSUME_ABSTRACT(SomeNamespace::SomeClass) BOOST_CLASS_TYPE_INFO(SomeNamespace::SomeClass, extended_type_info_no_rtti<SomeNamespace::SomeClass>) BOOST_CLASS_EXPORT(SomeNamespace::SomeClass) Can someone explain please? Thanks

3 2

Joining two forward ranges
by er 13 Jan '09

13 Jan '09

Hello, I'm looking to create an iterator which iterates over two joined ranges: [b0,e0)[b1,e1) I'm sure there's a way to do it with an iterator adapter. I was wondering, however, if someone knew of an already available implementation. Thank you.

2 2

boost[thread] in win32 we have API like WaitForMultipleObjects.How can we implement the same using boost thread
by Mrudul Mahurkar 13 Jan '09

13 Jan '09

Hi, i am working on boost::asio and boost thread library.I am creating client application that will work against any server.In that application am using threads and for their synchronisation i need the APIs like WaitForSingleObject and WaitForMultipleObject given by win32.so is there any API in boost ::thread library so that it can work like that. Mrudul Mahurkar Mumbai.India. Add more friends to your messenger and enjoy! Go to http://messenger.yahoo.com/invite/

3 2