January 2011 - Boost-users - lists.preview.boost.org

Spirit : Is possible to build rule step by step?
by Andrea Fontana 17 Jan '11

17 Jan '11

I'm trying to write a small example using spirit. I can write a rule in this way: my_rule = lit("example") | lit("example2"); Also in this way: std::string ex("example") std::string ex2("example2"); my_rule = lit(ex) | lit(ex2)) Is there a way to do this: std::string ex("example") std::string ex2("example2"); my_rule = lit(ex); my_rule = my_rule | lit(ex2)); I need to append strings from an array... Maybe I miss something from docs.

3 2

[Lambda] result_of support bug for placeholders?
by Michel MORIN 17 Jan '11

17 Jan '11

Hi, Let's we have using boost::lambda::_1; typedef boost::remove_reference<decltype(_1)>::type _1_type; Then, decltype(_1(1)) (and std::result_of<_1_type(int)>::type) is 'int const&', but boost::tr1_result_of<_1_type(int)>::type is 'int&". Is this a bug? Regards, Michel

1 0

Customize Message Queue
by derleader mail 16 Jan '11

16 Jan '11

Hi, I need advice how to customize Message Queue in boost / interprocess / ipc / message_queue . hpp > I need to insert additional argument in mq . send (& number , sizeof ( i ), 0 , ); and here mq . receive (& number , sizeof ( number ), recvd_size , priority, ); Its something like 2 messages in one queue. How I can do this? Regards Peter

2 1

Re: [Boost-users] functors created by boost::bind and move semantic
by Marat Abrarov 16 Jan '11

16 Jan '11

To: michi7x7 > > Marat Abrarov: > > Does anybody know why functional objects created by means of > > boost::bind have no move constructors. It would be rather useful > > especially when such functional objects consist of some smart pointers > > (like boost::shared_ptr). > > For example, move constructor of boost::shared_ptr is very light > > (comparing to its copy constructor) - it doesn't use any CAS at all. > > May be someone know the reason of move constructor absence or maybe > > someone know any plans of implementing it. > michi7x7: > I wondered about that too. The reason might be the huge ammounts of code > that need to be changed within boost::function. > However, I'm sure it will come soon. I mean not boost::function but functors created by boost::bind (boost::bind_t....) - they are rather simple - like tuples, so, I think, their move constructors would be simple too. But it would be great to have move constructors in boost::function too (which is more difficult because of current boost::function implementation). Interesting, what is about std::function (or std::tr1::function) in STL implementations with C++0x support, for example, in STL shipped with MS Visual C++ 2010. Marat Abrarov.

1 0

TTI library in the Boost sandbox
by Edward Diener 16 Jan '11

16 Jan '11

I have uploaded to the Boost sandbox 'tti' directory version 1.0 of the TTI library. The TTI library, which is an abbreviation for the Type Traits Introspection library, allows a programmer to introspect at compile time the inner elements of a C++ type. The introspection process depends on specifying the name of the inner element by different macros for different types of elements, and then using a generated metafunction to determine whether that element exists within the enclosing type. The inner elements which can be introspected are type, class template, member data, member function, static member data, and static member function. The TTI library is based on the type_traits_ext portion of the Concept Traits Library, with improvements and additions, and also lifts functionality, for the purposes of orthogonality, from Boost.MPL regarding introspection of types and templates. The purpose of the library is to provide a consistent set of interfaces for doing compile-time introspection which other template metaprogrammers can use in their code. There is a readme.txt in the top-level sandbox tti directory, for anyone browsing the sandbox for interesting libraries, which replicates what is written here. There is a build.txt file in the doc subdirectory for building the documentation and running the tests. The documentation is also included as part of the sandbox files for those who can not build the docs. There is a list of todo items in the sandbox on which I will be working for the next release, but the library is already feature rich. The library has been tested and works with gcc 4.30, 4.4.0, 4.5.0-1, 4.5.2, and VC++ 8.0, 9.0, and 10.0. It may also work with other compilers. There is a bug in versions of gcc prior to version 4 for which I will attempt to find a workaround if it is deemed important, which prevents it from working with the library. Questions, comments, suggestions, and bug reports are all welcome.

1 0

[Proto] Difficult Arity Check
by Hossein Haeri 16 Jan '11

16 Jan '11

Dear all, I have a syntax like ((_1 >> &f1 >> &f2) || (_2 >> &f3) || (_3 >> &f4 >> &f5 >> &f6)) >> g for which I already have a small grammar which along with my context do their job fine. The above expression is to act like g(f2(f1(x)), f3(y), f6(f5(f4(z)))) once invoked with _1 == x, _2 == y, and _3 == z. I am currently trying to add a safety layer on top of that which can emit a compile-error -- at the appropriate corner of my code -- that can prevent pass of functions with wrong arities. For example, g above needs to be a ternary. Although my context will notice it if g is not a ternary and produce a meaningful error message, that's too late. My poor user needs to be informed right at the instantiation time. So, I thought I'm going to use transforms to get my grammar manage that. To that end, I thought the _state of my grammar needs to be an MPL pair of an mpl::int_ and an mpl::bool_. Below is my algorithm with some plain English explanation. Please would someone help me for I found it utterly formulate that in Proto. 1) when<terminal<_>, mpl::pair<mpl::int_<1>, mpl::true_> > -- so far, so easy. 2) When logical_or<Arity, Arity>, return summation of the two arities, paired with the logical and of success value in the two branches: mpl::pair < mpl::plus < typename _state(EmtnArity(_left))::first, typename _state(EmtnArity(_right))::first >, mpl::and_ < typename _state(EmtnArity(_left))::second, typename _state(EmtnArity(_right))::second > > (This doesn't compile.) 3) When shift_right<Arity, Arity>, return mpl::int_<1> again, but, this time, paired with whether or not the right arity has been passed around logically anded with the success value so far: mpl::pair < mpl::int_<1>, mpl::and_ < typename _state(EmtnArity(_left))::second, is_callable_with_args(***) > > And, I'm totally stuck in the use of is_callable_with_args! Any helps please? Feel free to pass a completely new design too. TIA, --Hossein

2 3

Splitting Tiff RGBA into RGB and alpha data ...
by Corinna Kinchin 15 Jan '11

15 Jan '11

Hi there I have an uncompressed RGBA Tiff file (RGB + alpha channel) which I'm trying to split into two data files, one containing the RGB data, and the other containing the alpha (transparency) data. Here's a code snippet showing my (probably misguided) attempt at doing this: image_read_info< tiff_tag > info = read_image_info(in, tiff_tag()); if ( info._photometric_interpretation == 2 && info._samples_per_pixel == 4 ) { // RGBA rgba8_image_t rgba_img; // Read RGBA image ... read_image(in, rgba_img, tiff_tag()); // Set up planar image (to extract alpha channel) and rgb image // (to extract rgb data) rgba8_planar_image_t dst_planar_image(rgba_img.dimensions()); rgb8_image_t dst_rgb_image(rgba_img.dimensions() ); // Convert pixels to appropriate view copy_and_convert_pixels(view(rgba_img), view(dst_rgb_image)); copy_and_convert_pixels(view(rgba_img), view(dst_planar_image)); // Write out data ofstream out((ofilename+"rbg").c_str(), ios_base::binary), alphaout((ofilename +"alpha").c_str(), ios_base::binary); out.write(reinterpret_cast<const char*>(interleaved_view_get_raw_data(view(dst_rgb_image))), info._width * info._height * 3); alphaout.write(reinterpret_cast<const char*>(planar_view_get_raw_data(view(dst_planar_image), 3)), info._width * info._height); ... The alpha transparency data is ok as far as I can see, but the RGB data is mangled ... presumably I've misunderstood how the copy_and_convert_pixels() works. I'm new to the boost::gil library, and am finding the documentation a little hard to get to grips with, so any hints you could give would be greatly appreciated. Thanks a million in advance! All the best, ... Corinna

2 1

functors created by boost::bind and move semantic
by Marat Abrarov 15 Jan '11

15 Jan '11

Hi, Boost Users. Does anybody know why functional objects created by means of boost::bind have no move constructors. It would be rather useful especially when such functional objects consist of some smart pointers (like boost::shared_ptr). For example, move constructor of boost::shared_ptr is very light (comparing to its copy constructor) - it doesn't use any CAS at all. May be someone know the reason of move constructor absence or maybe someone know any plans of implementing it. Abrarov Marat.

2 1

Call C++ function from second Limux daemon
by derleader mail 15 Jan '11

15 Jan '11

Hi, I'm using boost for interprocess communication using message queue. There are 2 Linux daemons which must exchange data. In daemon A I want to call a C++ function from daemon B and pass some arguments for execution. The function in daemon B must execute the arguments and return the result to daemon A. Is it possible to implement this? Regards Peter

1 0

[Function Types] can_be_called
by Hossein Haeri 14 Jan '11

14 Jan '11

Dear all, May I please ask the people in charge of function_type to address: https://svn.boost.org/trac/boost/ticket/3783 I need it in my code and need to know when it's going to become available. TIA, --Hossein

2 1

[Proto] Anything Callable
by Hossein Haeri 14 Jan '11

14 Jan '11

Dear all, How can I have a grammar that matches a terminal of any callable type? Be it a function, function object, or a Boost.Function (specifically). TIA, --Hossein

3 6

[version] boost version check, header (.hpp) vs binaries (libs)
by alfC 14 Jan '11

14 Jan '11

Hi, I work in at least 3 computers (home/work/cluster, in which the cluster is really a shared filesystem between different architectures, etc). Besides that, for some projects I use the default version of boost in that system and in others a locally (user space) installed boost (that is, ranging from boost 1.33 to boost 1.46svn). Therefore is not uncommon that I mess the compilation lines and end up including headers from a certain version an linking to a different one. Hopefully this fails at link time: e.g. undefined reference to `boost::filesystem3::path::extension() const' but it is confusing and potentially it could compile (and then crash unexpectedly) I am aware of 'boost/version.hpp' but this is only for header verification. The question is, is there a Boost mechanism to verify that the header and the library in boost are consistent? (Hopefully at compile/link time) Thank you, Alfredo

2 1

Error '__Unwind_Resume' when trying to link codeblocks to boost::thread
by Bruno Ferraz do Amaral 14 Jan '11

14 Jan '11

Hi, I'm a new user of Boost Libraries, and I'd use multithreading linking the boost::thread class into my projetct. I'm using Code::Blocks, and I've built the Boost in Windows, and then I followed a wiki guide to link the libraries of the Boost.Thread with my Project in CodeBlocks. However I found this reference errors: C:\boost_1_45_0\stage\lib\libboost_thread-mgw45-mt-1_45.a(thread.o):thread.cpp:(.text+0x7d)||undefined reference to `__Unwind_Resume'| C:\boost_1_45_0\stage\lib\libboost_thread-mgw45-mt-1_45.a(thread.o):thread.cpp:(.text+0x66f)||undefined reference to `__Unwind_Resume'| C:\boost_1_45_0\stage\lib\libboost_thread-mgw45-mt-1_45.a(thread.o):thread.cpp:(.text+0x771)||undefined reference to `__Unwind_Resume'| C:\boost_1_45_0\stage\lib\libboost_thread-mgw45-mt-1_45.a(thread.o):thread.cpp:(.text+0xdbe)||undefined reference to `__Unwind_Resume'| C:\boost_1_45_0\stage\lib\libboost_thread-mgw45-mt-1_45.a(thread.o):thread.cpp:(.text+0xe53)||undefined reference to `__Unwind_Resume'| C:\boost_1_45_0\stage\lib\libboost_thread-mgw45-mt-1_45.a(thread.o):thread.cpp:(.eh_frame+0x12)||undefined reference to `___gxx_personality_v0'| ||=== Build finished: 6 errors, 0 warnings ===| What I have to do? I tried a lot of tutorials and guides in Google, but no one works...

4 5

[Serialization][Variant] serialization of variant containing boost::blank
by dariomt＠gmail.com 14 Jan '11

14 Jan '11

Hi, In the following code I show how a boost::variant that has a boost::blank cannot be serialized, and a proposed fix. I argue that this should work out of the box. This is with Boost 1.42.0. Am I doing something wrong here, or is serialization of boost::blank somewhere already? Thanks #include <boost/variant/variant.hpp> #include <boost/serialization/variant.hpp> #include <boost/archive/text_oarchive.hpp> #include <sstream> #ifdef FIX // suggested fix: non-intrusive serialize() for boost::blank namespace boost { namespace serialization { template<class Archive> void serialize(Archive &, boost::blank &, const unsigned int) { // Nothing to do here } } } #endif int main() { boost::variant<int, double> v1 = 666; boost::variant<boost::blank, int, double> v2 = 666; std::ostringstream os; boost::archive::text_oarchive oa(os); // works! oa << v1; // compile error about boost::blank not having a serialize method // #define FIX to compile oa << v2; return 0; }

1 0

ASIO: What is the proper way to upload a LARGE file?
by Ramon F Herrera 13 Jan '11

13 Jan '11

I have successfully been using this code to upload small stuff: try { // open tcp connection tcp::iostream net_stream(strSubmitServer, "12345"); // send file name net_stream << strFile << endl; // copy file char ch; while ((ch = fgetc(f)) != EOF) net_stream << ch; } catch (std::exception& e) { std::cout << "Exception: " << e.what() << std::endl; } Needless to say, the code above is extremely slow/inefficient for bulky transfers. Now I need to upload some large fails, and tried this: tcp::iostream net_stream(SERVER_ADDR, "12345"); // send file name net_stream << IN_FILE << "\n"; // send tags data size net_stream << count << "\n"; // copy tags data net_stream.write((char *)data, count); <<== app crashes here I figured that TCP would break the delivery into packets... TIA, -Ramon

3 2

Implications of using Boost with different versions of Visual Studio in the same computer?
by Ramon F Herrera 13 Jan '11

13 Jan '11

I have installed MSVS-2005 and MSVS-2010 in my development PC. I noticed that contrary to other libraries, every boost installation seems to be specific for: - The Boost version (duh!) - The MSVC Version For instance: libboost_system-vc80-mt-1_43.lib I would prefer to have all the -vc80- and -vc100- library files in the same folder, but perhaps I am looking for trouble? Then there is the issue of the includes. If such strategy turns out to be ill-advised, I would like to somehow "hide" my MSVS-2010 (don't want to uninstall it!) and then install Boost, built bye the VS-2005 toolchain. Thanks for your kind assistance. -RFH ps: Boost rules!! :-)

5 5

Serialization Boost 1.45
by Andrei Popa 13 Jan '11

13 Jan '11

Hi, I was building my project until recently with Boost 1.42. I have upgraded my Boost Version to 1.45. First of all, I see that now by default is used the dynamic linking to the Boost/Serialization. How can I set my code/project to link it statically to the Boost/Serialization? Second, I can not open my old files, I used only binary archives. I read about fix_six, but when I run it with my files I got the message "library version not equal to six"? How is this possible? How can I found out what "library version " do I have? What can I do to fix this problem? Thank you very much in advance! Andrei.

2 1

Boost 1.42 TR1 With GCC 4.5 selects wrong overload
by Abir Basak 13 Jan '11

13 Jan '11

I am using the portable style boost headers for tr1 implementation from Boost 1.42 such as #include <boost/tr1/type_traits.hpp> When i use gcc implementation of tr1 (Mingw GCC 4.5.1 from TDM, BOOST_HAS_GCC_TR1 defined) everything works fine. However when i select boost tr1 implementation (i.e not define the above macro) some stdlib functions such as std::abs gets wrong overloaded selected. For. e.g. in gcc/tr1 std::abs((signed char)5) returns an int while in boost/tr1 it returns a double. A compilable example is like #include <cstdlib> #include <iostream> #include <typeinfo> #include <boost/tr1/type_traits.hpp> int main(){ std::cout<<typeid(std::abs((short)5)).name(); //prints i with gcc/tr1 while prints d with boost/tr1 } it happens with type_traits header (may be few other also) but not with all tr1 headers(such as array). Thank you. abir basak

2 1

Boost 1.45 Serialization : Getting archive exception while loading
by Tejaswin Macharla 13 Jan '11

13 Jan '11

Hi, I have changed the sample code given in the Boost site for Serialization and the saving of the file works fine. The code works fine without the switch case as it executes save and load in the same session. I can save and load in the same session. If I close the application and open the application again and try to load the saved(serialized) file, it is throwing an Archive Exception. I am trying this for a game that I am working on so as to save and load in different sessions. Code: ----------------------------------------------------------------------- #include <fstream> // include headers that implement a archive in simple text format #include <boost/archive/text_oarchive.hpp> #include <boost/archive/text_iarchive.hpp> #include <boost/archive/binary_iarchive.hpp> #include <boost/archive/binary_oarchive.hpp> ///////////////////////////////////////////////////////////// // gps coordinate // // illustrates serialization for a simple type // class gps_position { private: friend class boost::serialization::access; // When the class Archive corresponds to an output archive, the // & operator is defined similar to <<. Likewise, when the class Archive // is a type of input archive the & operator is defined similar to >>. template<class Archive> void serialize(Archive & ar, const unsigned int version) { ar & degrees; ar & minutes; ar & seconds; } int degrees; int minutes; float seconds; public: gps_position(){}; gps_position(int d, int m, float s) : degrees(d), minutes(m), seconds(s) {} void gps_display() { std::cout << "\n Displaying....\n"; std::cout << degrees << " "<< minutes << " "<< seconds<<std::endl; } }; int main() { // create and open a character archive for output std::ofstream ofs("filename"); // create class instance const gps_position g(35, 59, 24.567f); gps_position newg; char ch; std::cout << "Enter an option\n 1. Save\n 2. Load\n 3. Display\n\n"; std::cin >> ch; switch(ch) { case '1': { // save data to archive boost::archive::text_oarchive oa(ofs); // write class instance to archive oa << g; // archive and stream closed when destructors are called } break; case '2': { // ... some time later restore the class instance to its orginal state // create and open an archive for input std::ifstream ifs("filename", std::ios::binary); boost::archive::text_iarchive ia(ifs); // read class state from archive ia >> newg; // archive and stream closed when destructors are called } break; case '3': newg.gps_display(); break; default: std::cout << "\nInvalid option\n"; } return 0; } ----------------------------------------------------------------------- Can we load a serialized file from the same program in a different session? -- Thanks & Regards ------------------------------ Tejaswin G Macharla -----------------------------

2 2

Portability on XBox, PS3, and iPhone
by Hossein Haeri 12 Jan '11

12 Jan '11

Dear all, 10 years ago, it was crucially important to know whether the Boost stuff one was about to use was compatible with the targeted C++ compiler. The Boost documentation was quite informative on that IIRC. With the recent shifts of the industry, the equivalent and more update information needs to be on portability of each Boost library on the modern platforms. Is there any chance I can get some sort of comparative table on that? As the subject-line suggests, I am specifically interested in XBox, PS3, and iPhone. TIA, --Hossein

4 10

[Proto] Terminal of Whatever is Derived from T
by Hossein Haeri 12 Jan '11

12 Jan '11

Dear all, How can I have a grammar which matches terminals of whatever type derived from T? TIA, --Hossein

2 2

Re: [Boost-users] boost::interprocess boost::interprocess xsi_shared_memory
by Daniel Neuberger 12 Jan '11

12 Jan '11

Is there a reason for not finishing the implementation of XSI (System V) shared memory? It seems like most of the work is already done. If there is no reason not to, I might be able to finish the implementation so that it can be included in the next release. I know that POSIX may be sufficient for some, but System V allows for much better control using SELinux since it lives in kernel space. FYI: This is in response to http://thread.gmane.org/gmane.comp.lib.boost.user/64537. Thoughts? Thanks. - Daniel

4 7

memory leak when using shared_from_this
by jochemspek 12 Jan '11

12 Jan '11

hi there, I'm having trouble understanding the cause of a memory leak apparently caused by using 'shared_from_this'. The following is the case: I have a AbstractProperty class with XMLNode as subclass. there exist two typedefs for these classes, namely typedef shared_ptr< XMLNode > XMLNodePtr and similar for AbstractProperty. in the XMLNode class I have the member function 'addChild' that adds a child to the XMLNode instance. I keep a reference to the parent by calling chid->setParent(), *with a shared pointer to this*. This results in a (quite severe) memory leak. I can resolve the problem by using a shared_ptr with a null-deleter ( see f.e. http://lists.boost.org/boost-users/2008/05/36281.php ), but I don't quite understand *why* the leak occurs, as I understand it so that the shared_from_this is a weak_ptr to this. I would much appreciate any comments. thanks, Jochem //---------------------------begin code----------------------------- const XMLNodePtr XMLNode::addChild( const string name ){ XMLNodePtr child( new XMLNode() ); child->setParent( static_pointer_cast< XMLNode, AbstractProperty >( shared_from_this() ) ); m_children.push_back( child ); return( child ); } //---------------------------end code----------------------------- -- View this message in context: http://boost.2283326.n4.nabble.com/memory-leak-when-using-shared-from-this-… Sent from the Boost - Users mailing list archive at Nabble.com.

1 1

Linking Error on WaitForSingleObject
by Anish Chandak 12 Jan '11

12 Jan '11

Hello All, I have a very unusual linking issue. It might be trivial but I could not find any reference after searching on the mailing list and other forums. I wrote a very simple example code (attached at the end of message) for using boost lock and mutex. The same piece of code compiles fine without linking error in one Visual Studio Project and gives the following linking error in another Visual Studio Project. I have made sure that the VS settings for each project match exactly. ERROR: error LNK2019: unresolved external symbol __imp__WaitForSingleObject__USE_VCR_MODE@8 referenced in function "public: void __thiscall boost::detail::basic_timed_mutex::lock(void)" (?lock@basic_timed_mutex@detail@boost@@QAEXXZ) I am using Visual Studio 2010 and boost 1_45_0. The code is linking to libboost_thread-vc100-mt-s-1_45.lib in both the projects. Any suggestions on what could be the possible cause of this or how to fix this will be highly appreciated. Thanks, Anish boost::mutex active_source_mutex; class TESTCLASS { public: void LOOP(int count) { int tcount = 0; for(int j=0; j<count; j++) { for(int k=0; k<10000000; k++) { tcount++; } printf("%d %d\n",j,tcount); } } void THREAD_CALLBACK() { boost::mutex::scoped_lock slock(active_source_mutex); LOOP(10); } }; int main() { TESTCLASS jk; boost::thread thread_1 = boost::thread(&TESTCLASS::THREAD_CALLBACK, &jk); boost::thread thread_2 = boost::thread(&TESTCLASS::THREAD_CALLBACK, &jk); thread_1.join(); thread_2.join(); return 0; }

3 3

[Proto] Non-const First Argument for Nested eval<>'s operator() ()
by Hossein Haeri 12 Jan '11

12 Jan '11

Dear all, In the Proto documentation, is there any reason why in all the specialisations of eval<Expr> for contexts, the first argument to their operator() is of type Expr& as opposed to const Expr&? TIA, --Hossein

2 2