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

Re: [Boost-users] Boost-users post acknowledgement
by Yong Qin 11 Aug '11

11 Aug '11

for VC, I try to link to: libboost_test_exec_monitor-vc100-mt-1_47.lib and got error: LNK1561: entry point must be defined VCApp_Test line 0 C/C++ Problem for MinGW, I try to link to: libboost_test_exec_monitor-mgw46-mt-1_47.a and got error: undefined reference to `__chkstk_ms' MinGWApp_Test line 0, external location: D:\boost_1_47_0-MinGW\stage\static\lib\libboost_unit_test_framework-mgw46-mt-1_47.a(unit_test_parameters.o):unit_test_parameters.cpp C/C++ Problem undefined reference to `std::__detail::_List_node_base::_M_hook(std::__detail::_List_node_base*)' MinGWApp_Test line 0, external location: D:\boost_1_47_0-MinGW\stage\static\lib\libboost_unit_test_framework-mgw46-mt-1_47.a(unit_test_suite.o):unit_test_suite.cpp C/C++ Problem undefined reference to `std::__detail::_List_node_base::_M_unhook()' MinGWApp_Test line 0, external location: D:\boost_1_47_0-MinGW\stage\static\lib\libboost_unit_test_framework-mgw46-mt-1_47.a(unit_test_suite.o):unit_test_suite.cpp C/C++ Problem

1 0

question about Boost naming
by Yong Qin 11 Aug '11

11 Aug '11

I am using Boost test library now and find there are several libraries there: libboost_unit_test_framework-mgw46-1_47.a libboost_unit_test_framework-mgw46-mt-1_47.a libboost_unit_test_framework-mgw46-s-1_47.a what are differences between the 3 libraies (with "s", or with "mt", or no "mt" and "s")? also, for VC++ compiler, there is: boost_unit_test_framework-vc100-mt-1_47.dll but MinGW have: libboost_unit_test_framework-mgw46-mt-1_47.dll libboost_unit_test_framework-mgw46-1_47.dll Why MinGW have more?

6 6

[boost-build] Dynamic linking dependency propagation on mac os x
by Tim Odenthal 11 Aug '11

11 Aug '11

Dear list, i'm trying to compile an existing project which compiles fine on Windows (where we link statically) and linux (where linking is default, i.e. dynamic) on Mac OS X (10.7). I'm essentially trying to use the same configuration as on linux, since darwin ~ gcc… The project makes use of (and therefore links to) several boost libraries including system, filesystem, serialization,… We create several libraries which depend upon each other: - lib1depends on serialization,filesystem,… - lib2 depends on lib1 - lib3 depends on lib1 and lib2 -> here's the first problem: the (implicit) dependence on system, filesystem and serialization is not automatically resolved anymore. At least, the linker complains about missing symbols, since the "g++"-command-line doesn't include the libraries. I could solve that by adding those libraries to the requirements with <use>/boost/serialization, etc. -lib4 depends on lib3 -> here's the real problem: again, the implicit dependence on (specifically) lib1 and the boost-libs is lost. I can use the same trick as before with the boost-libs, but whatever I do, I cannot make the "lib1.dylib" appear in the "g++" command-line. When I execute the command by hand - adding lib1.dylib - it executes fine & the linker does not complain. How/why does that happen? My user-config.jam simply consists of: using darwin : : : <compileflags>-DBOOST_WAVE_SUPPORT_THREADING=0 ; What else do I need to configure differently on Mac OS X? Thanks a lot in advance for any suggestions Tim

1 0

Windows Installer Build
by asif saeed 11 Aug '11

11 Aug '11

Hi, I installed Boost 1.47 using Windows Installer. I have Visual C++ 2010 Express installed and I installed Platform SDK for 2003 and Windows Driver Kit, that give me MFC/ATL, etc., separately. However, I have to set the directories in each project before I build anything otherwise the VC++ 2010 picks up the Windows SDK 7.0 that gets installed by default by Visual C++ 2010 Express. So Boost installer, I suspect, too picked up the wrong SDK, Windows SDK 7, directories instead of the those of the platform SDK. I suspect so because upon including ASIO header and just getting an IP Address string, I get build warnings like these: Please define _WIN32_WINNT or _WIN32_WINDOWS appropriately. For example: - add -D_WIN32_WINNT=0x0501 to the compiler command line; or - add _WIN32_WINNT=0x0501 to your project's Preprocessor Definitions. Assuming _WIN32_WINNT=0x0501 (i.e. Windows XP target). File0001.cpp Please define _WIN32_WINNT or _WIN32_WINDOWS appropriately. For example: - add -D_WIN32_WINNT=0x0501 to the compiler command line; or - add _WIN32_WINNT=0x0501 to your project's Preprocessor Definitions. Assuming _WIN32_WINNT=0x0501 (i.e. Windows XP target). File0002.cpp Second_test.cpp Please define _WIN32_WINNT or _WIN32_WINDOWS appropriately. For example: - add -D_WIN32_WINNT=0x0501 to the compiler command line; or - add _WIN32_WINNT=0x0501 to your project's Preprocessor Definitions. Assuming _WIN32_WINNT=0x0501 (i.e. Windows XP target). Any idea how I can fix this? Best, Asif

1 0

Re: [Boost-users] [asio] user defined composed operations and strands
by Marat Abrarov 11 Aug '11

11 Aug '11

Hi, Tacheon. > If async_write_data is called and io_service::run is called from multiple threads, strands are should > be necessary to avoid raced. > The question is: How do I support handlers wrapped in a strand and how do I get this strand forwarded > to async_read? > As far as I understand this should be somehow possible by using asio_handler_invoke but the > documentation is a bit too fuzzy on the topic for me to understand. There is no such info in Asio docs. You may look at asio-samples (http://sourceforge.net/projects/asio-samples) There is ma::context_wrapped_handler class (and ma::make_context_wrapped_handler helper function) that makes such a wrapper for intermediate handler. But you shouldn't use > typedef boost::function<void (error_t const &)> completion_handler_t; because it hides asio_hanlder_invoke defined for the original (user supplied) handler. You may write something like this: class io_port { public: //Ctors and so on... template<typename Handler> void async_read_data ( raw_buffer_t & buffer , Handler const & handler) { //uses a composed operation itself... boost::asio::async_read ( socket , boost::asio::buffer(header) , ma::make_context_wrapped_handler(handler, boost::bind ( &io_port::handle_header_received<Handler> , this , boost::ref(buffer) , handler , boost::asio::placeholders::error))); } //first intermediate handler template<typename Handler> void handle_header_received ( raw_buffer_t & data , Handler const & handler , boost::system::error_code const & error { if(error) { handler(error_t(error)); } else { prepare_buffer(data, header.perparation_data()); //pseudo-code //uses a composed operation itself... boost::asio::async_read ( socket , boost::asio::buffer(data) , ma::make_context_wrapped_handler(handler, boost::bind(&io_port::handle_data_received<Handler> , this , boost::ref(buffer) , handler , boost::asio::placeholders::error) ); } } //second intermediate handler template<typename Handler> void handle_data_received ( raw_buffer_t & data , Handler const & handler , boost::system::error_code const & error ) { if(error) { handler(error_t(error)); } else { handler(error_t()); } } }; Also, it will be very helpful to read Asio sources: /boost/asio/impl/read.hpp: 339 and 145 (Boost 1.47). May be it' will be better to implement io_port class similar to boost::asio::detail::read_op. Regards, Marat Abrarov.

2 1

question about changing permission with boost
by rudy b 11 Aug '11

11 Aug '11

Hi, I am trying to copy a file from one folder to another, but I am also trying to access to the last_write_time of the file, BEFORE copying it to the other folder, And I notice, whenever I am inside a Read-only directory, and trying to access the "last_write_time" of a Read-only file, I get the following error message: what(): boost::filesystem::last_write_time: Permission denied: "src/ temp1/methodologies.ppt" In this scenario "temp1" is a Read-only directory, also methodologies.ppt is a Read-only file as well. Does anyone know how to get around this? Is there a way to change the directory/file permission (to write) prior to copying, and then change the permission back again? I need to access the time-stamp before copying the file, because right after I copy the file, the date-stamp will be updated. I would really appreciate some help, please advise, thanks, RB

1 0

[BGL] bundled properties and forward declarations (changes from 1.42 -> 1.46.1 ?)
by Anders Wallin 10 Aug '11

10 Aug '11

I have code using bundled properties roughly like this that compiles with boost 1.42: struct EdgeProperty; typedef adjacency_list< ... , EdgeProperty , ...> MyGraph; typedef graph_traits<MyGraph>::edge_descriptor Edge; struct EdgeProperty { Edge next; }; When trying to compile this with boost 1.46.1 I get "error: forward declaration of EdgeProperty’". I am not sure why. The obvious solution would be to move the definition of EdgeProperty before the MyGraph typedef, but then I run into a circular dependency problem as I want to store an edge_descriptor in EdgeProperty. any ideas? Anders

2 3

[bimap] Problem with finding elements in a vector_of
by Teto 10 Aug '11

10 Aug '11

Hi, I've a peculiar bimap defined as such: bbi::bimap< bbi::unordered_set_of< TFullId>, bbi::vector_of<CBindDescriptor>> And I want to find all the TFullId matching a certain CBindDescriptor. but I can't succeed in compiling. Here are 10 lines of code: http://boost.codepad.org/99V38S9U Thx for any hint matt

2 3

[BGL] resize/reserve on graph?
by Geoff Hilton 10 Aug '11

10 Aug '11

If I know that during the execution of the algorithm the graph size must be grown by X number of vertices in one go, is there a way this can be done to avoid multiple allocations such as when using std::vector's reserve(..) or resize(..) member functions without subverting the BGL's documented interface? Currently I use a for loop which iterates over add_vertex(vertex_property, graph). Can a reserve and/or resize equivalent otherwise please be added somewhere? Thank you, Geoff

3 6

installing boost_1_47_0
by rudy b 10 Aug '11

10 Aug '11

Hi, I am newb in using boost. And, I am having so much trouble installing and running a simple example. I would REALLY appreciate some help. I tried to install boost_1_47_0.tar.gz on my unix machine, and I followed the stpes http://www.boost.org/doc/libs/1_47_0/more/getting_started/unix-variants.html to first build, and install the libraries. I had no problem running some the examples (provided on the link above, that was using some of its already built libraries), and they all worked. The problem is when I try to run an example that is using a boost_filesystem. I tried uninstalling/installing different boost versions, cause I thought the error may be version dependent, but it wasn't. I also tried to built the entire library using different versions of gcc (3.4.6, 4.3.5, 4.4.0, 4.5.0, 4.5.1, 4.5.2, 4.6.0), and I get the same error messages. I even made sure that when compile, I have the boost_1_4x_0 path in my compile script. I also made sure that I have appended boost_1_4x_0 path to my $path. I even tried to have -lboost_filesystem in my compile script, but this didn't help either !!! This is the type of error that I am getting: /tmp/cczDoDOU.o: In function `__static_initialization_and_destruction_0(int, int)': sync.cpp:(.text+0xe03): undefined reference to `boost::system::generic_category()' sync.cpp:(.text+0xe1e): undefined reference to `boost::system::generic_category()' sync.cpp:(.text+0xe39): undefined reference to `boost::system::system_category()' /tmp/cczDoDOU.o: In function `boost::filesystem3::last_write_time(boost::filesystem3::path const&, long)': And, when I change the version of compiler, I get something like: /tmp/ccw1Sepi.s: Assembler messages: /tmp/ccw1Sepi.s:408: Error: unknown pseudo-op: `.cfi_personality' /tmp/ccw1Sepi.s:409: Error: unknown pseudo-op: `.cfi_lsda' /tmp/ccw1Sepi.s:531: Error: unknown pseudo-op: `.cfi_personality' /tmp/ccw1Sepi.s:532: Error: unknown pseudo-op: `.cfi_lsda' /tmp/ccw1Sepi.s:722: Error: unknown pseudo-op: `.cfi_personality' /tmp/ccw1Sepi.s:723: Error: unknown pseudo-op: `.cfi_lsda' /tmp/ccw1Sepi.s:1046: Error: unknown pseudo-op: `.cfi_personality' Please advise, thanks, RB

3 3