August 2013 - Boost-users - lists.preview.boost.org

[units] "Simple" conversion from Kelvin to Celsius
by Michael Powell 22 Aug '13

22 Aug '13

Hello, It's been a couple of years since I've done any work with Boost.Units, and I am finding that I want to make a dimensionally-safe conversion from Kelvin to Celsius. I've got an I2C controller that reports in Kelvin, but I want to report it to the end-user in Celsius. The examples I am finding seem to "almost" fit the bill, but I am having a difficult time extrapolating a sane conversion. First, would help if I understood what the Boost.Units SI (my assumed base-system) base-temperature units were. I gather Fahrenheit? Or is it indeed Kelvin? I could be mistaken. Thank you... Regards, Michael Powell

4 4

Trouble building 1.5.4 on OS X 10.6
by Jo Meder 22 Aug '13

22 Aug '13

Hi, I'm having trouble building 1.5.4 on OS X 10.6 for Intel only Universal builds (32 and 64 bit). This is my command line for b2: ./b2 macosx-version=10.6 macosx-version-min=10.6 threading=multi architecture=x86 address-model=32_64 I've also tried a simpler command line just specifying the architecture and address model, same results. Originally I was having trouble building all the libraries, in particular the thread lib wasn't building. Now mysteriously it is. However the libraries are being built with the ppc and ppc64 architectures as well as i386 and x86_64, which doesn't seem right given the options I'm specifying. You can see the unwanted archs in the build lines too, for example: "g++-4.0" -ftemplate-depth-128 -arch x86_64 -arch i386 -arch ppc64 -arch ppc -O3 -finline-functions -Wno-inline -Wall -gdwarf-2 -fexceptions -Wno-long-double -isysroot /Developer/SDKs/MacOSX10.6.sdk -mmacosx-version-min=10.6 -arch i386 -arch x86_64 -DBOOST_ALL_NO_LIB=1 -DNDEBUG -I"." -c -o "bin.v2/libs/context/build/darwin-4.0/release/address-model-32_64/architecture-x86/link-static/macosx-version-min-10.6/macosx-version-10.6/threading-multi/unsupported.o" "libs/context/src/unsupported.cpp" There are 4 arch flags near the start for x86_64, i386, ppc64 and ppc. There are also two more for x86_64 and i386 further on, which I assume are the ones the darwin.jam file has added based on my options passed to b2. I'm guessing these build lines are only being shown when there's a problem as there certainly isn't one for every file, but I assume they're representative. Now the libs seem to be building but I'm getting link errors from libboost_thread: Undefined symbols: "boost::thread_detail::enter_once_region(boost::once_flag&)", referenced from: boost::detail::set_current_thread_data(boost::detail::thread_data_base*)in libboost_thread.a(thread.o) boost::detail::get_current_thread_data() in libboost_thread.a(thread.o) _thread_proxy in libboost_thread.a(thread.o) _thread_proxy in libboost_thread.a(thread.o) boost::(anonymous namespace)::make_external_thread_data() in libboost_thread.a(thread.o) "boost::thread_detail::commit_once_region(boost::once_flag&)", referenced from: boost::detail::set_current_thread_data(boost::detail::thread_data_base*)in libboost_thread.a(thread.o) boost::detail::get_current_thread_data() in libboost_thread.a(thread.o) _thread_proxy in libboost_thread.a(thread.o) _thread_proxy in libboost_thread.a(thread.o) boost::(anonymous namespace)::make_external_thread_data() in libboost_thread.a(thread.o) "boost::thread_detail::rollback_once_region(boost::once_flag&)", referenced from: boost::detail::set_current_thread_data(boost::detail::thread_data_base*)in libboost_thread.a(thread.o) boost::detail::get_current_thread_data() in libboost_thread.a(thread.o) _thread_proxy in libboost_thread.a(thread.o) _thread_proxy in libboost_thread.a(thread.o) boost::(anonymous namespace)::make_external_thread_data() in libboost_thread.a(thread.o) ld: symbol(s) not found collect2: ld returned 1 exit status I saw there was a problem building the thread lib on OS X so I've added: #define BOOST_THREAD_DONT_USE_ATOMIC to boost/config/user.hpp. That seems related because in the headers that appears to control which file with the enter_once_region and so on functions gets included. If anyone has any insight I'd appreciate it. Regards, Jo Meder

1 0

memory used by vector<int> vs vector<char>
by Lloyd 21 Aug '13

21 Aug '13

Hi, If I create a vector<int> of 100 elements and a vector<char> of 100 elements, will there be any difference in the total memory used by them? or which is more efficient? Thanks, Lloyd

3 3

Identify cycles in a undirected graph
by NsPx 20 Aug '13

20 Aug '13

Hi all, I use undirected_dfs to detect all cycles in a undirected graph. I detect a cycle with DFSVisitor.back_edge(e,g) but how can I get the list of vertices defining this cycle ? Thanks by advance

4 4

[log] Retrieving the current trivial severity level
by Chris Stylianou 20 Aug '13

20 Aug '13

Is there any method of retrieve the current severity level from a "boost::log::sources::severity_logger_mt<boost::log::trivial::severity_level>" logger. I would like to store the current severity level in a temporary variable, upgrade it to a new level using "BOOST_LOG_SEV(logger::get(), boost::log::trivial::trace)" to write a trace-type message, and then downgrade the logger back to the previous setting by recalling "BOOST_LOG_SEV(logger::get(), temporaryVariable)" -- View this message in context: http://boost.2283326.n4.nabble.com/log-Retrieving-the-current-trivial-sever… Sent from the Boost - Users mailing list archive at Nabble.com.

2 1

Re: [Boost-users] [rational] numeric_limits not specialized
by Daryle Walker 20 Aug '13

20 Aug '13

Ticket 9017 added: <https://svn.boost.org/trac/boost/ticket/9017>. If you want, you can visit that web page to register yourself to get updates. I’ll try to get at it (and the other reports) in the near-future, after adding some bug reports of my own, and my current project. Daryle W. Sent from Windows Mail From: Michael Olea Sent: ‎Monday‎, ‎August‎ ‎19‎, ‎2013 ‎7‎:‎15‎ ‎PM To: boost-users(a)lists.boost.org Hi. I have some template code that needs slightly different behavior when it is instantiated with exact types like int or boost::rational rather than when it is instantiated with floating point types. The compile-time test depends on std::numeric_limits<T>::is_exact. I was surprised when this came back false for boost::rational until I found that std::numeric_limits<boost::rational<long>>::is_specialized is false. Shouldn't boost::rational specialize numeric_limits? I'm using boost 1.54.0 with clang: Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn) compiling for C++1. _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

[thread] Can't compile once.cpp for ppc in 1.54
by James Walker 20 Aug '13

20 Aug '13

Attempting to build the 1.54.0 version of the thread library for ppc on Mac OS X 10.6.8, using Apple's GCC 4.0, I get this error: boost_1_54_0/boost/atomic/detail/gcc-ppc.hpp: In member function 'bool boost::atomics::detail::base_atomic<T, int, 4u, Sign>::compare_exchange_strong(T&, T, boost::memory_order, boost::memory_order) volatile [with T = unsigned int, bool Sign = false]': /Volumes/Home/Mercurial/boost_1_54_0/boost/atomic/detail/gcc-ppc.hpp:1190: error: impossible constraint in 'asm' If I use GCC 4.2, I don't get the error. But I'm hesitant to use that compiler, because it usually produces code that won't run on G3's. -- James W. Walker, Innoventive Software LLC <http://www.frameforge3d.com/>

2 3

[rational] numeric_limits not specialized
by Michael Olea 19 Aug '13

19 Aug '13

Hi. I have some template code that needs slightly different behavior when it is instantiated with exact types like int or boost::rational rather than when it is instantiated with floating point types. The compile-time test depends on std::numeric_limits<T>::is_exact. I was surprised when this came back false for boost::rational until I found that std::numeric_limits<boost::rational<long>>::is_specialized is false. Shouldn't boost::rational specialize numeric_limits? I'm using boost 1.54.0 with clang: Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn) compiling for C++1.

1 0

newbie question about boost and OpenSSL 1.0.1e (boost won't compile with it)
by michael＠skydogsoftware.com 19 Aug '13

19 Aug '13

Hello, I have a Windows C++ application that uses Boost 1_48 and OpenSSL 0.9. I just upgraded to OpenSSL 1.0.1e, and now boost won't compile, e.g. I get the following errors: 14>c:\_ot\svn\ERB\_bin\3rd\include\boost/asio/ssl/impl/context.ipp(411) : error C2039: 'PEM_read_bio_DHparams' : is not a member of '`global namespace'' 14>c:\_ot\svn\ERB\_bin\3rd\include\boost/asio/ssl/impl/context.ipp(411) : error C3861: 'PEM_read_bio_DHparams': identifier not found 14>c:\_ot\svn\ERB\_bin\3rd\include\boost/asio/ssl/impl/context.ipp(421) : error C2039: 'DH_free' : is not a member of '`global namespace'' 14>c:\_ot\svn\ERB\_bin\3rd\include\boost/asio/ssl/impl/context.ipp(421) : error C3861: 'DH_free': identifier not found Will upgrading to the latest version of boost correct this? Thanks in advance for any answers/help/tips. -Michael Compton michael(a)skydogsoftware.com

2 1

[spirit] Parsing the following question
by Michael Powell 19 Aug '13

19 Aug '13

Hello, I would like to parse the following: "{0,0}" with the following constraints: rule<iterator> r %= '{' >> ( int_(0) | int_(1) | int_(2) )[phx::ref(m_mux) = _1] >> ',' >> ( int_(0x29) | int_(0x39) | int_(0x49) )[phx::ref(m_addr) = _1] >> '}'; So that m_mux is capturing either 0, 1, or 2, and m_addr is capturing either 29H, 39H, or 49H (in decimal). Having some difficulty though with the following errors: Description Resource Path Location Type static assertion failed: error_invalid_expression detector line 176, external location: /home/mpowell/srcb/Dchem/SEEKERe/dchem-seekere/tools/installed/boost/include/boost/spirit/home/qi/nonterminal/rule.hpp C/C++ Problem Maybe I don't need the int_ at all? Thank you... Regards, Michael Powell

1 1