November 2010 - Boost-users - lists.preview.boost.org

[MPI] MPI_Issend() support
by Riccardo Murri 23 Nov '10

23 Nov '10

Hello, I read in the manual that there's no equivalent of MPI_Issend() (noblocking synchronous send) in Boost.MPI. Is there any architectural obstacle to implementing it, or is it just that no-one has asked for it yet? Best regards, Riccardo

2 2

Boost.MPI library binary
by Hicham Mouline 22 Nov '10

22 Nov '10

Hello if I find a library binary of Boost.MPI, does it mean it must have been built against 1 of the many MPI implementations? Or is there some sort of link-level compatibility ?

3 3

boost.mpi: some MPI implementation that mixes mpi process from both windows and linux
by Hicham Mouline 22 Nov '10

22 Nov '10

Hello, Does such an implementation exist? and can it be used with Boost.MPI? Failing that, I am looking at having the GUI of my application in windows and the mpi processes running on a set of linux boxes I have. The question is: How can the GUI (not an mpi process) control the mpi environment (the bunch of mpi processes that run on the linux boxes)? regards,

1 0

Solaris Sun Studio 12, boost 1.44 - choosing version of STL
by Paulo Moura Guedes 22 Nov '10

22 Nov '10

Hi, I want to use the standard Solaris STL libCstd, so I'm using the "stdlib=native" switch in bjam. However, stlport4 is being picked up. How can I force boost not to use stlport4? Thanks in advance, Paulo

3 2

Linker error on windows
by Lars Mogensen 22 Nov '10

22 Nov '10

Hi, I am using BOOST to do math and have not been able to get the "findroot_demo.cpp" example of root finding to work. There seem to be a reference to rint(), which is not implemented in the C++ library for windows. How can I solve this - my understanding of BOOST is that it is cross-platform? 1>------ Build started: Project: boostfindzero, Configuration: Debug Win32 ------ 1>Compiling... 1>findroot_demo.cpp 1>Linking... 1>findroot_demo.obj : error LNK2019: unresolved external symbol _rint referenced in function "public: static double __cdecl boost::numeric::interval_lib::detail::x86_rounding::to_int(double const &)" (?to_int@x86_rounding@detail@interval_lib@numeric@boost@@SANABN@Z) 1>D:\dokumenter\lvm_AM\Visual Studio 2008\Projects\boostfindzero\Debug\boostfindzero.exe : fatal error LNK1120: 1 unresolved externals 1>Build log was saved at "file://d:\dokumenter\lvm_AM\Visual Studio 2008\Projects\boostfindzero\boostfindzero\Debug\BuildLog.htm" 1>boostfindzero - 2 error(s), 0 warning(s) ========== Build: 0 succeeded, 1 failed, 0 up-to-date, 0 skipped ========== Thanks in advance. /Lars

4 8

boost::Asio async_accept not throwing exception
by Lloyd Kanakkassery 22 Nov '10

22 Nov '10

Hi, I have modified the sample program provided in the asio samples, just to accept a connection like this void TCPServer::StartAccept() { try { TCPConnection::Pointer NewConnection=TCPConnection::Create(Acceptor.io_service()); Acceptor.async_accept(NewConnection->socket(),boost::bind(&TCPServer::HandleAccept,this, NewConnection,boost::asio::placeholders::error)); } catch(std::exception& e) { cout<<"exception"<<endl; } } Another program is already listening on the port "13" (A service version of the same program). When this new program tries to listen (async_accept) on port "13", no exception is thrown!. How can I check weather the port is already open? Thanks, Lloyd

3 5

boost::asio socket as class member
by Claude 22 Nov '10

22 Nov '10

Hi! I want to put a boost::asio::ip::tcp::socket as a private member of my class and then create the object in one of the class methods. I actually use this code: class myClass{ private: boost::asio::io_service io_service; boost::asio::ip::tcp::socket s; }; But this code don't work (don't compile). Why? -- View this message in context: http://boost.2283326.n4.nabble.com/boost-asio-socket-as-class-member-tp3047… Sent from the Boost - Users mailing list archive at Nabble.com.

3 3

problem of opcode not supported on mipsel for Boost_1_39_0
by Yan Wu 22 Nov '10

22 Nov '10

Hi, I got an error when building OpenOffice.org on mipsel platform, the error is following: Compiling: comphelper/source/misc/logging.cxx {standard input}: Assembler messages: {standard input}:101: Error: opcode not supported on this processor: mips1 (mips1) `ll $3,8($4)' {standard input}:103: Error: opcode not supported on this processor: mips1 (mips1) `sc $2,8($4)' {standard input}:141: Error: opcode not supported on this processor: mips1 (mips1) `ll $3,4($4)' {standard input}:143: Error: opcode not supported on this processor: mips1 (mips1) `sc $2,4($4)' dmake: Error code 1, while making '../../unxlngmips.pro/slo/logging.obj' The <boost/shared_ptr.hpp>, Boost_1_39_0, is used in the above logging.cxx. Is this a bug for the smart_ptr on mipsel platform ? I found there was a similar bug 2972(https://svn.boost.org/trac/boost/ticket/2972). But I notice that the fix for 2972 is not integrated to the current milestone yet. Right? Where can I find the fix for this bug ? Best Regards, Yan

2 2

[MPI] lifetime of data sent through communicator::isend()
by Riccardo Murri 21 Nov '10

21 Nov '10

Hello, I'm using mpi::communicator::isend where the payload data comes from a pointer (e.g., "req = comm_.isend(dest, tag. *data)"). When am I free to "delete" the pointer? Immediately after isend() returns? Or should I wait until the request is completed? Does setting BOOST_MPI_HOMOGENEOUS have any influence on this? Best regards, Riccardo

4 5

[proto] Unable to display_expr a spirit expression
by Marcin Zalewski 21 Nov '10

21 Nov '10

Hello. I have a following program: #include <boost/spirit/include/qi.hpp> #include <boost/spirit/include/phoenix_core.hpp> #include <boost/proto/core.hpp> int main () { // boost::proto::display_expr(boost::phoenix::arg_names::arg1); boost::proto::display_expr(boost::spirit::int_[boost::phoenix::arg_names::arg1]); } That works as expected: subscript( terminal(N5boost6spirit3tag4int_E) , terminal(N5boost7phoenix5actorINS0_8argumentILi0EEEEE) ) When I include <boost/spirit/home/phoenix/operator.hpp> (as the last include), however, the output of the program changes to: subscript( terminal(N5boost6spirit3tag4int_E) , terminal() I am not sure if this is expected, unsupported, or a bug. Anyone? Just to put this in context, I am trying to figure out how does spirit intercept phoenix expressions (e.g., in semantic actions) to wrap them in proto terminals. The infrastructure of spirit is rather complex, so if anyone can point me in the right direction, I would really appreciate it. -m

2 2