- Boost-users - lists.boost.org

How to deduce (possibly void) return type of member function?
by Terry Golubiewski 23 May '10

23 May '10

If I have a class... struct X { int f(double x); void f(int y); }; ... and a template Y that wants to deduce the return-type of some member functions of T... template<class T> struct Y { typedef typename what_goes_here<T>::type f_double_return_type; typedef typename what_goes here<T>::type f_int_return_type; }; I"ve tried using... typedef BOOST_TYPEOF_TPL(static_cast<X*>(0)->f(double(0))) f_double_type; // this works ... but ... typedef BOOST_TYPEOF_TPL(static_cast<X*>(0)->f(int(0)) f_int_type; // does not work for "void" Of course, I wouldn't know which one returns "void" or not ahead of time. terry

2 5

[profile_templates] Unable to run the template profiler
by Nathan Crookston 22 May '10

22 May '10

All, A template metaprogram I've written consumes too much memory when compiling on VC8. To debug this, I'm trying to use Steven Watanabe's template profiler. Unfortunately, I've been unable to run it successfully on my code. After some casting about I found [1] which resolved some initial problems, it seems. As a sanity check, I tried the following: ncrookston: ~/boost/tools/profile_templates/test $ bjam toolset=msvc error: Unable to find file or target named error: '/boost//regex' error: referred from project at error: '../src I'm definitely not a bjam/boost-build expert. Does anyone have suggestions on how I can fix this? My platform: XP 32-bit using VIsual Studio 2005 (VC8) Thanks, Nate [1] <http://lists.boost.org/Archives/boost/2009/05/152016.php>

3 3

Re: [Boost-users] [msm] submachine interaction
by Christophe Henry 22 May '10

22 May '10

> Hello - > > I've created a state machine in MSM successfully, but cannot figure > out how to interact with a submachine state object. > > Using the tutorial machine in "Defining a submachine", can a playlist > be provided to Playing? I'd like to call a constructor, or a 2 stage > init function of Playing, but haven't figure out how. > > I could use the "FSM" parameter of the on_entry method of the > submachine to pull the playlist, but it appears that FSM is the > submachine itself (Playing), not the containing machine (Player) as it > is for regular states. > > Thanks for you help, > Gordon Smith Hi Gordon, You could call "get_state<player_::Playing&>()" on the player instance but I agree that the current behavior makes little sense, so I just changed it. The "FSM" parameter of on_entry (and on_exit) for a submachine now will be the containing machine (player), so that submachines' entry and exit conditions get the same parameters as substates. Of course, substates of the submachine still get as FSM parameter the submachine, not the containing machine. As we are still far from the release, we still have time to make this change. Should I have broken somebody's code, please shout ;-) I now moved MSM to the trunk, and all changes are now done there, so you need to get: https://svn.boost.org/svn/boost/trunk/boost/msm/back/state_machine.hpp (it is also in the sandbox, just in case). Thanks for the report, Christophe

1 0

Need example of substitution of priority queue
by Eric Fowler 22 May '10

22 May '10

I wish to run a breadth_first_search() on a graph where I am defining the priority queue (I have a hand-rolled sort order). Is there any reason I cannot use std::priority_queue<> to do this? Does anyone have an example, or know of one in the docs? Thanks Eric

2 5

Re: [Boost-users] boost/numeric/ublas/matrix_sparse.hpp does not play well with RICE
by John O. Woods 22 May '10

22 May '10

I have partially resolved the issue. I deleted all boost libraries and headers from my machine and clean re-compiled and installed 1.43. I can now compile with unordered_map successfully. I suspect the compiler was using one version of the libraries and a different version of the headers. Unfortunately, I still cannot #include <boost/numeric/ublas/matrix_sparse.hpp> while also including rice/Data_Type and rice/Constructor. The #undefs Steven recommended don't appear to make a difference. I've included the current code below. It compiles perfectly if I comment the first line. If I replace the Data_Type.hpp header with a different rice header, it still doesn't work -- I've tried Class.hpp, String.hpp, Object.hpp, and Constructor.hpp. FYI, here's the address for rice: http://rice.rubyforge.org/main.html Thanks again, John #include <rice/Data_Type.hpp> #include <boost/numeric/ublas/matrix_sparse.hpp> int main() { return 0; } $ make g++ -I. -I. -I/usr/lib/ruby/1.8/i486-linux -I. -D_FILE_OFFSET_BITS=64 -I/usr/lib/ruby/gems/1.8/gems/rice-1.3.2/ruby/lib/include -fPIC -fno-strict-aliasing -g -g -O2 -fPIC -Wall -g -c main.cpp In file included from /usr/local/include/boost/numeric/ublas/storage_sparse.hpp:23, from /usr/local/include/boost/numeric/ublas/vector_sparse.hpp:16, from /usr/local/include/boost/numeric/ublas/matrix_sparse.hpp:16, from main.cpp:2: /usr/local/include/boost/numeric/ublas/storage.hpp:60: error: expected primary-expression before ‘)’ token /usr/local/include/boost/numeric/ublas/storage.hpp:65: error: expected primary-expression before ‘)’ token /usr/local/include/boost/numeric/ublas/storage.hpp:79: error: expected primary-expression before ‘)’ token /usr/local/include/boost/numeric/ublas/storage.hpp: In member function ‘typename ALLOC::size_type boost::numeric::ublas::unbounded_array<T, ALLOC>::max_size() const’: /usr/local/include/boost/numeric/ublas/storage.hpp:174: error: expected primary-expression before ‘)’ token /usr/local/include/boost/numeric/ublas/storage.hpp:174: error: expected ‘;’ before ‘ruby_xmalloc’ /usr/local/include/boost/numeric/ublas/storage.hpp:174: error: expected primary-expression before ‘)’ token /usr/local/include/boost/numeric/ublas/storage.hpp: In member function ‘typename ALLOC::size_type boost::numeric::ublas::bounded_array<T, N, ALLOC>::max_size() const’: /usr/local/include/boost/numeric/ublas/storage.hpp:357: error: expected primary-expression before ‘)’ token /usr/local/include/boost/numeric/ublas/storage.hpp:357: error: expected ‘;’ before ‘ruby_xmalloc’ /usr/local/include/boost/numeric/ublas/storage.hpp:357: error: expected primary-expression before ‘)’ token In file included from /usr/local/include/boost/numeric/ublas/vector_sparse.hpp:16, from /usr/local/include/boost/numeric/ublas/matrix_sparse.hpp:16, from main.cpp:2: /usr/local/include/boost/numeric/ublas/storage_sparse.hpp: At global scope: /usr/local/include/boost/numeric/ublas/storage_sparse.hpp:240: error: expected primary-expression before ‘)’ token make: *** [main.o] Error 1

4 3

Inconsistent with variant with bool
by Robin 22 May '10

22 May '10

Hi I'm facing a inconsistent behavior with Variant when bool is defined. The below code will return i as 0 typedef boost::variant <int> VariantType; VariantType variant; int i = boost::get <int> (variant)); // default to 0 If I set a bool in the variant typedef boost::variant <bool, int> VariantType; VariantType variant; int i = boost::get <int> (variant)); // this throws fail get Is it intended behavior? Robin

3 2

Sentinel vertex?
by Eric Fowler 22 May '10

22 May '10

Does BGL define a sentinel value that is always invalid as a vertex? I am thinking of something analogous to the null char at the end of ASCIIZ strings. Eric

2 1

Boost.MPI 1.43.0 and openmpi 1.4.2 problems
by Benjamin Behringer 22 May '10

22 May '10

Hi, I'm new to boost and openmpi. I'm on a Mac installed the newest version of openmpi through "port" and compiled boost by myself. There is also an openmpi version coming with my Mac in version 1.2.8. I tried to compile and run a small sample but ran into different problems. So the following example won't work for me with openmpi 1.4.2 (through port) but with openmpi 1.2.8: #include <boost/mpi.hpp> #include <iostream> #include <string> #include <boost/serialization/string.hpp> namespace mpi = boost::mpi; int main(int argc, char* argv[]) { mpi::environment env(argc, argv); mpi::communicator world; if (world.rank() == 0) { world.send(1, 0, std::string("Hello")); std::string msg; world.recv(1, 1, msg); std::cout << msg << "!" << std::endl; } else { std::string msg; world.recv(0, 0, msg); std::cout << msg << ", "; std::cout.flush(); world.send(0, 1, std::string("world")); } return 0; } I compiled with (1.4.2): openmpic++ boost_mpi_test.cpp -lboost_mpi -lboost_serialization -o boost_mpi_test Running that test says: openmpirun -np 2 boost_mpi_test *** The MPI_Pack_size() function was called before MPI_INIT was invoked. *** This is disallowed by the MPI standard. *** Your MPI job will now abort. [pillepalleben:54081] Abort before MPI_INIT completed successfully; not able to guarantee that all other processes were killed! *** The MPI_Pack_size() function was called before MPI_INIT was invoked. *** This is disallowed by the MPI standard. *** Your MPI job will now abort. [pillepalleben:54082] Abort before MPI_INIT completed successfully; not able to guarantee that all other processes were killed! Compiling with (1.2.8): mpic++ boost_mpi_test.cpp -lboost_mpi -lboost_serialization -o boost_mpi_test While the test runs fine: mpirun -np 2 boost_mpi_test Hello, world! So shall I use 1.2.8? Why I'm running in those issues? Thanks in advance, ben

1 0

boost::unit_test main function problem
by Tamas Nagy 22 May '10

22 May '10

Hello! I have a test.cpp file: #define BOOST_TEST_DYN_LINK #define BOOST_TEST_MODULE Hello // nothing It contains nothing, but the 2 defines. I have a long Makefile of a large project. I copied the Makefile of the large project to the _unittest dir. I modified the paths in the Makefile to the proper paths. The large project has a main.cpp, which has a main() function. If i want to compile with the boost::unit_test, actually the test.cpp, it says: In file included from ../_unittest/src/test.cpp:31: ../_prog/src/main.cpp: In function ‘int main(int, char**)’: ../_prog/src/main.cpp:603 error: redefinition of ‘int main(int, char**)’ /usr/include/boost/test/unit_test.hpp:57: error: ‘int main(int, char**)’ previously defined here make: *** [../_unittest/src/test.o] Error 1 So the problem is, that there are 2 entry points. main.cpp cannot be modified to contain only the functions to be tested, because a large project relies on it. So the ideal solution would be let the original project code untouched, create a test binary that tests the functions from the large project. I'm a beginner the boost::test, can you help me solve this problem? Tamas

1 0

where to get Boost.Log?
by Kenny Riddile 22 May '10

22 May '10

I know it isn't part of the official distribution yet, but where can I get the latest usable version of Boost.Log?

2 1