- Boost-users - lists.boost.org

async_read_some() not found??
by Claude 12 Aug '11

12 Aug '11

I have this code (from async_read_until boost documentation): boost::asio::streambuf b; void handler(const boost::system::error_code& e, std::size_t size) { if (!e) { std::istream is(&b); std::string line; std::getline(is, line); // ... } } boost::asio::async_read_until(s, b, "\r\n", handler); But I obtain this error on Ubuntu 11.04 (boost 1.45.0): boost/asio/impl/read_until.hpp|388|*error: request for member ‘async_read_some’* in ‘((boost::asio::detail::read_until_delim_op<boost::asio::basic_stream_socket<boost::asio::ip::tcp>*, std::allocator<char>, void (*)(const boost::system::error_code&, unsigned int)>*)this)->boost::asio::detail::read_until_delim_op<boost::asio::basic_stream_socket<boost::asio::ip::tcp>*, std::allocator<char>, void (*)(const boost::system::error_code&, unsigned int)>::stream_’, which is of non-class type ‘boost::asio::basic_stream_socket<boost::asio::ip::tcp>*’ I Included in my project: libpthread libboost_system -- View this message in context: http://boost.2283326.n4.nabble.com/async-read-some-not-found-tp3732368p3732… Sent from the Boost - Users mailing list archive at Nabble.com.

2 2

[filesystem] misidentifies file type on Solaris x86 64-bit
by Will Mason 11 Aug '11

11 Aug '11

Hello, I'm using Boost 1.46.1 with filesystem v3. Code that I've been running successfully on Windows, Linux and FreeBSD started producing very mysterious results on Solaris. I boiled it down to this: boost::filesystem will not identify a directory as a directory. Here's a test program: #include <boost/filesystem.hpp> int main() { try { boost::filesystem::path d(boost::filesystem::absolute("dir")); if (boost::filesystem::exists(d)) { std::cout << d << " already exists" << std::endl; } else { boost::filesystem::create_directory(d); std::cout << d << " is " << (boost::filesystem::is_directory(d) ? "" : "not") << " a directory" << std::endl; } } catch (std::exception& e) { std::cout << "Exception: " << e.what() << std::endl; } return 0; } And here is a snippet of a terminal session compiling and running the program: $ CC -o fstest fstest.cpp -library=stlport4 -Iexternal/solaris/x86/64/suncc/5.11/include external/solaris/x86/64/suncc/5.11/lib/libboost_filesystem.a external/solaris/x86/64/suncc/5.11/lib/libboost_system.a $ ./fstest "/export/home/will/Projects/cap/dir" is not a directory $ file /export/home/will/Projects/cap/dir /export/home/will/Projects/cap/dir: directory $ Thanks for any help, Will

1 0

Boost Unit test on Eclipse
by Yong Qin 11 Aug '11

11 Aug '11

I have tried to use boost unit test on Eclipse CDT w/ MinGW and VC++. But I can not link to the static boost library: libboost_unit_test_framework-vc100-mt-1_47.lib. The code I used: #defineBOOST_TEST_MAIN #defineBOOST_TEST_MODULEMasterTestSuite #include<boost/test/unit_test.hpp> BOOST_AUTO_TEST_SUITE(FooTests) BOOST_AUTO_TEST_CASE(ctor_test1) { Fooa; BOOST_CHECK_EQUAL(a.name(),MY_PROJECT_NAMESPACE::Foo::DEF_NAME); } BOOST_AUTO_TEST_CASE(ctor_test2) { conststd::stringtheName("Blah"); Fooa(theName); BOOST_CHECK_EQUAL(a.name(),theName); } BOOST_AUTO_TEST_SUITE_END() What is wrong?

4 4

program_options: accepting NaNs for doubles
by Rhys Ulerich 11 Aug '11

11 Aug '11

Hi all, program_options does not accept NaN or Inf as valid double input. For example, the following sequence modifies the "Getting Started" first.cpp [1] to accept double input: cp $BOOST_SOURCE/libs/program_options/example/first.cpp . perl -pi -e 's/<int>/<double>/;' first.cpp LDFLAGS="-lboost_program_options" make first # ...or whatever The modified example indeed barfs on NaN or Inf input: $ ./first --compression=5.5 Compression level was set to 5.5. $ ./first --compression=Inf error: in option 'compression': invalid option value 'Inf' $ ./first --compression=NaN error: in option 'compression': invalid option value 'NaN' Two questions: 1) Does this count as a bug? 'std::cin >> foo' doesn't grok Inf or NaNs (on my system anyhow) but methods like strtod() do. Users providing explicit infinities or not-a-numbers as options parameters likely don't do so accidentally. This Inf/NaN behavior is not mentioned in the active tickets. 2) Can anyone suggest a workaround (shy of using string-based parameters and calling strtod myself)? Thanks, Rhys [1] http://www.boost.org/doc/libs/1_47_0/doc/html/program_options/tutorial.html…

2 2

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