October 2006 - Boost-users - lists.preview.boost.org

copy_graph gives excpetion
by gast128 17 Oct '06

17 Oct '06

Dear all, I get a crash when copying a graph. Probably something simple. Does anyone know? void Foo() { typedef boost::adjacency_list<boost::listS, boost::listS, boost::bidirectionalS, boost::property<boost::vertex_index_t, int> > Graph; Graph graph1; Graph graph2; FillGraph(&graph1); //normal copy boost::copy_graph(graph1, graph2); //<-- } Wkr, me

1 0

boost installation problem...
by rahul mehra 17 Oct '06

17 Oct '06

hi all, i have downloaded boost 1.33.1 from the following link. http://easynews.dl.sourceforge.net/sourceforge/boost/boost_1_33_1.zip i want to build the bjam file with the source but it is giving ne this error message:- [root@localhost jam_src]# sh ./build.sh : command not found : command not found : command not found '/build.sh: line 12: syntax error near unexpected token ` '/build.sh: line 12: `echo_run () i tried to run the configure script but it is also giving some error:- [root@localhost boost_1_33_1]# sh ./configure : command not found8: : command not found3: '/configure: line 25: syntax error near unexpected token `do '/configure: line 25: `do can you please help me in installing boost... thanks Rahul _________________________________________________________________ One and only Ash. Find out all about her. Only on MSN Search http://server1.msn.co.in/profile/aishwarya.asp

1 0

Question regarding template
by YYW 17 Oct '06

17 Oct '06

Hello all, why the codes below is wrong in g++ 3.4.2? an error occurs in the line: cout<<CMatrix<double>::eye(5); -------------------------------------- #include <ostream> #include <boost/numeric/ublas/matrix.hpp> #include <boost/numeric/ublas/io.hpp> using namespace boost::numeric::ublas; using namespace std; template <class T> class CMatrix { matrix<T>mat_; public: CMatrix(const identity_matrix<T>& m) { mat_ = m; } static CMatrix<T> eye(int size) { return identity_matrix<T>(size); } template <class T_> friend ostream& operator << (ostream& os, CMatrix<T_>& m); }; template<class T_> ostream& operator << (ostream& os, CMatrix<T_>& m) { os<<m.mat_; return os; } int main() { CMatrix<double>I = CMatrix<double>::eye(5); cout<<I; cout<<CMatrix<double>::eye(5);//this line ==> error } ---------------------------------------- Thanks in advance!

3 4

[boost][VC++ 7.1][link] Crash when linking in release
by Washburn 17 Oct '06

17 Oct '06

Hi all, I experience violent crashes trying to link Boost in release. Very simple example : - I build Boost using bjam (all standard config, tool msvc-71) - I create an empty MFC project with all defaults settings except for include directories ("C:\Boost\include\boost-1_33_1") and lib directories ("C:\Boost\lib\") - I try to execute this : boost::filesystem::path::default_name_check(boost::filesystem::native); boost::filesystem::path p; p = "c:\\"; The last line works as expected in Debug but crashes in release : Unhandled exception at 0x00402917 in testboost2.exe: 0xC0000005: Access violation reading location 0x00000004. It doesn't help trying to trace the code (int this case it's in m_path_append( ) but it depends on the code being tried) Any idea ? Thanks in advance !

1 0

Re: [Boost-users] [interprocess] Segment existance check
by Ion Gaztañaga 17 Oct '06

17 Oct '06

Hi, > I don't know if I'm just missing it or if this feature doesn't exist, but is > there a feature in interprocess to check that a segment exists prior to > attempting to find or construct a variable in that segment? If it doesn't > exist this would be an extremely useful feature so that when an application > goes to create an object in shared memory where the shared memory segment > has not been created the program doesn't just segfault but gives the > application at least a small chance of exiting gracefully (i.e. a thrown > exception). If you hold a reference to a managed_shared_memory the segment always exists, since in Interprocess, the processes is detached from the segment in the destructor of the managed_shared_memory. This is different from Shmem, where you could call "close()". No other processes can destroy the segment while you have a opened/created managed_shared_memory. At best, in UNIX, it will unlink it, so new processes won't be able to find it, but old processes attached to the segment will continue working fine. I don't know if I have responded correctly to your question, so could you show me an example of what do you want to achieve? Regards, Ion

1 0

Re: [Boost-users] [interprocess] trouble porting from boost.shmem
by Pataki Szilard 16 Oct '06

16 Oct '06

Hello Ion, > Thanks for the code. I'll try to see what's happening, but it seems a > bit strange. Which Interprocess version are you using? I'll recommend > you to test your code agains latest cvs code. Apart from that, can you > tell which system are you using, windows, or unix? And, can you which is > the value of SEGMENT_SIZE, SEGMENT_NAME, MAP_NAME you are using (it > shouldn't matter, but who knows...)? Sorry, I always forget these things: - linux, gcc 3.4.5, amd64 - I was using the snapshot of Interprocess from 2006.08.04 - I fetched Interprocess from CVS and it still crashes - I fetched the new Interprocess snapshot from 2006.10.13 and it doesn't compile, although it's the same code, which compiled with Interprocess 2006.08.04 and Interprocess.cvs. I'll send you the compiler errors soon. - const int SHMEM_SIZE = 50000000; - const char* SEGMENT_NAME = "/segment"; - const char* MAP_NAME = "mapDB"; - still works with shmem. Very weird. Regards, Szilard

2 1

storing some particular lambda expression with boost::function
by Levent Yilmaz 15 Oct '06

15 Oct '06

Hi, I have this construct: class foo {}; void fbar(foo*) {} int main() { using namespace boost::lambda; std::vector<foo> foos(3); std::for_each( foos.begin(), foos.end(), bind<void>(&fbar, &_1) ); } OK, this works. But I want to store the lambda expression, and wasn't able to to it. The following obvious try doesn't work: boost::function<void(foo*)> func = bind<void>(&fbar, &_1); std::for_each( foos.begin(), foos.end(), func ); How should I do this? thx, - Levent

1 0

boost::thread queries
by Rob Caldecott 15 Oct '06

15 Oct '06

I have just started to investigate the boost::thread library and have a couple of questions (Win32/VS2005): 1. I have an application that creates lots of relatively short-lived threads. When a thread is created I store the handle in an array, and at regular intervals, I remove completed threads from the array by checking the thread state via ::GetExitCodeThread. The application runs 24/7 so I don't want an array containing thousands of "dead" handles. I thought I'd replace this code with a boost::thread_group, but I cannot see any way to remove completed threads from the internal list. This means the std::list used by boost::thread_group would keep on increasing in size until the application exits, which I want to avoid. If I decide to roll my own thread group code, how can I check if a boost::thread is still active or not? 2. I am getting a memory leak when my application exits, which has already been mentioned here. Does anyone have a fix for this which I can apply to the 1.33.1 source (then rebuilding the thread lib by hand)? It would be much appreciated. I am also unsure if this leak will be fixed for 1.34 - does anyone know for certain? Thanks in advance.

8 12

[interprocess] trouble porting from boost.shmem
by Pataki Szilard 14 Oct '06

14 Oct '06

Hello! I'm having trouble porting my application from boost.shmem to boost.interprocess. Here are my typedefs: //-------------------- typedef std::pair< const int, MyNode> MyPair; typedef boost::interprocess::allocator < MyPair, boost::interprocess::managed_shared_memory::segment_manager> MyAlloc; typedef boost::interprocess::multimap< const int, MyNode, std::less<const int>, MyAlloc> MyMap; typedef MyMap::iterator MyMapIterator; //-------------------- In the first process I do: //-------------------- boost::interprocess::shared_memory_object::remove(SEGMENT_NAME); boost::interprocess::managed_shared_memory segment(boost::interprocess::create_only, SEGMENT_NAME, SEGMENT_SIZE); MyAlloc myAlloc(segment.get_segment_manager()); std::less<const int> myComparator; MyMap* myMap = segment.construct<MyMap>(MAP_NAME)(myComparator, myAlloc); //-------------------- In the second process I do: //-------------------- boost::interprocess::managed_shared_memory segment(boost::interprocess::open_only, SEGMENT_NAME); MyMap* myMap = segment.find<MyMap>(MAP_NAME).first; MyMapIterator it = myMap->begin(); // <- CRASHES HERE! for ( ; it != myMap->end; it++) { //... } //-------------------- The myMap pointer in the second process is not NULL, it has the same value as in the first process. Everything worked fine with boost.shmem. Anybody got a clue, what is wrong here and what is so different between how boost.shmem and boost.inteprocess perform these operations? Can boost.shmem and boost.interprocess coexist? If I'm stuck with boost.shmem for storing the map in shared memory, can I still use an upgradable mutex from boost.interprocess? Thanks, Szilard

2 1

Memory Footprint of BGL Nodes and Edges
by Wolfram Koska 14 Oct '06

14 Oct '06

Hi, I would like to know if there is any info on how much memory boost graph's Nodes and Edges consume (assuming listS)? I searched, but didn't find anything about this. I don't need exact values, a good estimate would be sufficient (to be able to predict the memory consumption of graphs with a huge number of nodes and edges). Thanks in advance, Wolfram Koska P.S.: While I'm at it, is there any native way to check if a node/edge is still in a graph?

2 1