December 2008 - Boost-users - lists.preview.boost.org

Boost.Asio 1.37 - can't compile with VC++ 2003 (version 1.35 is fine)
by lipik 29 Jan '09

29 Jan '09

Hi, I just upgraded my Boost distribution, and have immediately run into a compilation problem: both Intel and VC++ compilers complain that: - 'QueueUserAPC' : is not a member of 'operator``global namespace''' - 'QueueUserAPC': identifier not found, even with argument-dependent lookup 1.35 compiles fine. I checked and this API is available in Winbase.h. Can't figure out why it is not being found - there don't seem to be be any C++ version directives hiding it inside winbase. I am on XP, and _WIN32_WINNT is set to 0x0501, so this is not the culprit either. Pre-compiled headers are turned off, and I have defined WIN32_LEAN_AND_MEAN (to avoid winsock conflict). Any help greatly appreciated. Thanks Amit

2 1

boost::interprocess shared memory performance
by Andy Wiese 22 Jan '09

22 Jan '09

This is my first experience with using shared memory for anything more than trivial IPC. Thanks again to Ion Gaztañaga for getting the library working on FreeBSD. I'm developing initially on OS-X 10.5.6 with boost_trunk, but the eventual target platform is FreeBSD--I'm just an old mac programmer used to cushy development tools. I originally used managed_mapped_file, and got everything working, but I was disappointed by performance. Profiling with shark, I found that the application was spending a lot of time in msync, which according to man(2) is used to synchronize mapped memory with the filesystem. It makes sense to me that the interprocess containers I was creating could have a lot of file I/O overhead, and managed_mapped_file was probably not a good choice. So I rewrote the code to use managed_shared_memory instead of managed_mapped_file, thinking that it would eliminate the file I/O and therefore be faster. However, I am surprised that it is not much faster, and when I profile with Shark on os-x I see that it is still spending a lot of time in msync, specifically whenever a managed_shared_memory object is destroyed. (in boost::interprocess::mapped_region::flush(unsigned long, unsigned long), which is called within basic_managed_shared_memory's destructor) Does managed_shared_memory really need to call msync? I see that I should optimize my code to cache the managed_shared_memory objects so that fewer create/deletes are necessary, but this is still going to happen fairly frequently and I wonder if this expensive msync call is necessary. In the tradition of coder forums everywhere, someone will probably ask what I'm trying to accomplish and whether there may be a better way. Suggestions welcome. I'm writing a little cgi driven database utility that queries data stored in a filesystem directory using a simple query language. I would like to keep indexes of the data to speed query resolution. The utility is old-school cgi, so all its resources (such as indexes) have to be instantiated into memory each time the cgi process is started. I could write indexes to files, but then I incur a de/serialization overhead that is expensive. My intention was to keep the indexes as ready-to-use interprocess::maps in shared memory, to be used by all invocations of the cgi. It works, but the performance of the shared memory is poor enough that I'm not getting much increase over just doing a brute force search through the datafiles. All suggestions appreciated! Andy

4 5

Boost and Intel Compiler 11
by Matthieu Brucher 09 Jan '09

09 Jan '09

Hi, I've read on the archive that some work needed to be done to support Intel Compiler 11 (Linux). Was progress made in that matter ? Matthieu -- Information System Engineer, Ph.D. Website: http://matthieu-brucher.developpez.com/ Blogs: http://matt.eifelle.com and http://blog.developpez.com/?blog=92 LinkedIn: http://www.linkedin.com/in/matthieubrucher

4 15

Boost.Function in 1.37.0
by Amit Sonawane 09 Jan '09

09 Jan '09

Environment: WinXP SP2/Visual Studio Team System 2005/Boost 1.37.0 32-bit Using boost.function in Boost 1.37.0 generates several errors. Compiler extract : 28>c:\software\boost\boost_1_37_0\boost_1_37_0\boost\function\function_t emplate.hpp(80) : error C2143: syntax error : missing ';' before 'namespace' 28>c:\software\boost\boost_1_37_0\boost_1_37_0\boost\function\function_t emplate.hpp(980) : error C2838: '()' : illegal qualified name in member declaration 28> c:\software\boost\boost_1_37_0\boost_1_37_0\boost\function\function_temp late.hpp(80) : see reference to class template instantiation 'boost::function0<R>' being compiled 28>c:\software\boost\boost_1_37_0\boost_1_37_0\boost\function\function_t emplate.hpp(995) : error C2804: binary 'operator ==' has too many parameters 28>c:\software\boost\boost_1_37_0\boost_1_37_0\boost\function\function_t emplate.hpp(1002) : error C2804: binary 'operator !=' has too many parameters 28>c:\software\boost\boost_1_37_0\boost_1_37_0\boost\function\function_t emplate.hpp(1098) : error C2764: 'R' : template parameter not used or deducible in partial specialization 'boost::function<R(void)>' 28>c:\software\boost\boost_1_37_0\boost_1_37_0\boost\function\function_t emplate.hpp(1098) : error C3412: 'boost::function<R(void)>' : cannot specialize template in current scope 28>c:\software\boost\boost_1_37_0\boost_1_37_0\boost\function\function_t emplate.hpp(90) : error C2061: syntax error : identifier 'function_buffer' 28> c:\software\boost\boost_1_37_0\boost_1_37_0\boost\function\function_temp late.hpp(96) : see reference to class template instantiation 'boost::boost::detail::function::function_invoker1<FunctionPtr,R,T0>' being compiled 28>c:\software\boost\boost_1_37_0\boost_1_37_0\boost\function\function_t emplate.hpp(106) : error C2061: syntax error : identifier 'function_buffer' 28> c:\software\boost\boost_1_37_0\boost_1_37_0\boost\function\function_temp late.hpp(113) : see reference to class template instantiation 'boost::boost::detail::function::void_function_invoker1<FunctionPtr,R,T0 >' being compiled In Boost 1.36.0 the code complies ok. Is it boost function or my code. Esta mensagem pode conter informação confidencial e/ou privilegiada. Se você não for o destinatário ou a pessoa autorizada a receber esta mensagem, não deverá utilizar, copiar, alterar, divulgar a informação nela contida ou tomar qualquer ação baseada nessas informações. Se você recebeu esta mensagem por engano, por favor avise imediatamente o remetente, respondendo o e-mail e em seguida apague-o. Agradecemos sua cooperação. This message may contain confidential and/or privileged information. If you are not the addressee or authorized to receive this for the addressee, you must not use, copy, disclose, change, take any action based on this message or any information herein. If you have received this message in error, please advise the sender immediately by reply e-mail and delete this message. Thank you for your cooperation.

6 5

Updating my local boost.
by Robert Jones 06 Jan '09

06 Jan '09

I guess this question is more appropriate to the Subversion maillist, but I'm not subscribed to that, so it might take upto 24 hours to appear there. I'm hoping someone here will know the answer! My organisation maintains its boost libraries under its own Subversion repository. This strikes me as a bit odd, but no reason really why not AFAICS. However, the version in our repository is 1.33.1, and I'd like to upgrade to latest 1.36 release. Our network prevents me accessing the Boost subversion repository, so I have to download the tarball and unpack. Then what do I do? I basically want to overwrite our existing boost code with the newer code, but maintain our local edit history so we can see what's changed locally between our old boost code and new. Hence there are lots of additions, changes and deletions, all of which I want to be able see (that's see the differences between our old code and new code, not all the individual boost commits). How can I do this? Thanks - Rob.

5 5

[thread] locking and time
by Ernst Murnleitner 05 Jan '09

05 Jan '09

Hello, I have some questions about threads. It would be very nice if somebody could clarify. 1) recursive_timed_mutex: It is possible to use a time period for locking of a recursive_timed_mutex by using the approprioate timed_lock(). But there is not lock class with such a constructor? 2) (recursive timed) shared mutex: Is it not possible to lock with a period of time instead of ptime? 3) change of system time: What happens in a timed_lock if the system time is advanced by e.g. one hour? Is there a difference of using ptime or time_period? 4) read/writer mutex: if there are some recursive locks by one thread and this thread should become the writer, do I run in a deadlock? 5) What is the meaning of the various ..._lock_t (e.g. )? Is there a documentation? ad 1) There seems only to be a constructor with ABSOLUTE time: unique_lock(Lockable & m,boost::system_time const& abs_time) ad 2) There seems no possibility to use boost::posix_time::time_period for locking. ad 3) If the clock is set to one hour in the future. I think a timed_lock could take one hour at this moment? If yes, does it happen also if I give a time period? ad 4) if it is possible to recusively lock in one thread, it should be possible to make the thread a writtrer thread (exclusive owner) independently on how often it was locked before. Is this possible? ad 5) Only internal usage?: unique_lock(Lockable& m_,adopt_lock_t); unique_lock(Lockable& m_,defer_lock_t); unique_lock(Lockable& m_,try_to_lock_t) Thank you in advance, Greetings, Ernst

3 3

[BGL] vertex_index_t, edge_index_t?
by Geoff Hilton 05 Jan '09

05 Jan '09

Quick question: Given an arbitrary graph of type: typedef boost::adjacency_list< boost::listS, //out edge list selector boost::vecS, //vertex list selector boost::bidirectionalS, //directedness selector VertexPropertyWrapper_type, EdgePropertyWrapper_type> Graph_type; Where: typedef boost::property< boost::vertex_index_t, unsigned long long, MyVertexProperty> VertexPropertyWrapper_type; And: typedef boost::property< boost::edge_index_t, unsigned long long, MyEdgeProperty> EdgePropertyWrapper_type; How does setting edge/vertex_index_t as an interior property affect..anything? I saw in a post on this list somewhere a while ago that this was a necessary workaround for something or other (that which I don't recall now). In the post it actually said to use "int" (or was it unsigned int?) rather than unsigned long long but in any case I left it as unsigned long long and it hasn't affected anything negatively thus far so I left it that way. I believe I put it at unsigned long long because the type that was suggested for the *_index_t happened to match a weight type used in the post, but whether that was intentionally and implicitly or unintentionally and explicitly I don't remember. Thank you, Geoff

5 11

[BGL] Bundled properties and property maps
by Alex Brooks 05 Jan '09

05 Jan '09

Hi, I've used BGL before for representing graphs, but this is my first time with an algorithm. I've been stuck for many hours trying to define a property map with bundled properties, I'd really appreciate some help. I'm using gcc4.3, boost 1.34. An example program and the compiler error are: ------------------------------------------------------------- #include <boost/graph/graph_traits.hpp> #include <boost/graph/adjacency_list.hpp> #include <boost/graph/astar_search.hpp> using namespace std; class Point { public: double p_[2]; }; struct EdgeProp { double cost; }; typedef boost::adjacency_list<boost::vecS, boost::vecS, boost::undirectedS, Point*, EdgeProp> Graph; typedef boost::graph_traits<Graph>::vertex_descriptor VertexRef; class AStarHeuristic : public boost::astar_heuristic<Graph, double> { public: double operator()( VertexRef vertex ) const {return 0.0;} }; // exception for termination struct FoundGoal {}; // // A* Visitor that terminates when we reach the goal // class AStarVisitor : public boost::default_astar_visitor { public: AStarVisitor( const VertexRef &goal ) : goal_(goal) {} template <class Graph> void examine_vertex( VertexRef vertex, Graph &g ) { if( vertex == goal_ ) throw FoundGoal(); } private: VertexRef goal_; }; bool searchThroughGraph( const Graph &graph, VertexRef start, VertexRef goal, std::vector<VertexRef> &shortestPath ) { AStarHeuristic aStarHeuristic; AStarVisitor aStarVisitor( goal ); typedef boost::property_map<Graph, double EdgeProp::*>::type EdgeWeightMap; EdgeWeightMap edgeWeightMap = get(&EdgeProp::cost, graph); std::vector<VertexRef> predecessors(num_vertices(graph)); try { astar_search( graph, start, aStarHeuristic, boost::weight_map(edgeWeightMap). predecessor_map(&predecessors[0]). visitor(aStarVisitor) ); } catch ( const FoundGoal& ) { shortestPath.clear(); for ( VertexRef vertex = goal;; vertex = predecessors[vertex] ) { shortestPath.push_back( vertex ); if ( predecessors[vertex]==vertex ) break; } std::reverse( shortestPath.begin(), shortestPath.end() ); return true; } // If we get to here, we couldn't reach the goal return false; } int main() { Graph graph; std::vector<VertexRef> path; searchThroughGraph(graph,0,1,path); return 0; } ------------------------------------------------------------- /home/a.brooks/shh/src/libs/prm/test/stupid.cpp: In function ‘bool searchThroughGraph(const Graph&, VertexRef, VertexRef, std::vector<unsigned int, std::allocator<unsigned int> >&)’: /home/a.brooks/shh/src/libs/prm/test/stupid.cpp:64: error: conversion from ‘boost::bundle_property_map<const boost::adjacency_list<boost::vecS, boost::vecS, boost::undirectedS, Point*, EdgeProp, boost::no_property, boost::listS>, boost::detail::edge_desc_impl<boost::undirected_tag, unsigned int>, EdgeProp, const double>’ to non-scalar type ‘searchThroughGraph(const Graph&, VertexRef, VertexRef, std::vector<unsigned int, std::allocator<unsigned int> >&)::EdgeWeightMap’ requested ------------------------------------------------------------- Does anyone know what might be wrong? More generally, I'm a bit mystefied by the "<Graph, double EdgeProp::*>::type" syntax -- can anyone give me any pointers about what this syntax means, or where I can look it up? (googling for colon-colon-star doesn't help much). I also tried pattern-matching from the documentation page about bundled properties (the weight_map(get(&Highway::miles, map)) example) but couldn't get this to compile either. Is this example correct? Thanks very much in advance, Alex

4 19

[boost-users][mpl, enable_if] checking if a type has specific dependant type
by Igor R 01 Jan '09

01 Jan '09

Hello, I'd like to define a trait that would evaluate to true for types that have some specific dependant type (DependentType), to use like this: template<class Type> func(int i, typename boost::enable_if<HasDependentType<Type> >::type* dummy = 0) { } I tried this way: template<class Type, class Enable = void> struct HasDependentType : mpl::false_ {}; template<class Type> struct HasDependentType<Type, typename Type::DependentType> : mpl::true_ {}; ...but it appears incorrect (partial specialization is never ised). Thanks in advance.

2 2

[filesystem] windows program crashes in debug mode ( DbgUiConnectToDbg )
by Stephen 01 Jan '09

01 Jan '09

Trying to port a mac program onto windows. I've been having some trouble getting the boost portion of it to work. It appears to run fine in release mode*, but crashes when debugged. Below is an excised piece of code that seems to replicate the problem, and works perfectly in release mode but crashes when debugging. I'm using Code::Blocks under XP, with Boost 1.37 with Mingw 3.4.5 (linking to boost_system-mgw34-mt-1_37.lib and boost_filesystem-mgw34-mt-1_37.lib ) (Also: It doesn't crash when I run it in windows 95 compatibility mode, but that's not really an option because that completely breaks the program it's meant to be a part of). Code: #include "boost/filesystem/operations.hpp" #include "boost/filesystem/path.hpp" #include "boost/progress.hpp" #include <iostream> namespace fs = boost::filesystem; int main( int argc, char* argv[] ) { fs::path p("chapters"); fs::directory_iterator dir_iter(p), dir_end; std::cout<<"got here"; return 0; } Outputs "got here" in release, just crashes in debug, with backtrace: (gdb) backtrace #0 0x7c90120f in ntdll!DbgUiConnectToDbg () from C:\WINDOWS\system32\ntdll.dll #1 0x7c96c201 in ntdll!RtlpNtMakeTemporaryKey () from C:\WINDOWS\system32\ntdll.dll #2 0x0022f724 in ?? () #3 0x7c96c63e in ntdll!RtlpNtMakeTemporaryKey () from C:\WINDOWS\system32\ntdll.dll #4 0x00448270 in __gnu_cxx::__mt_alloc<wchar_t>::_S_binmap () #5 0x003e0000 in ?? () #6 0x00448278 in __gnu_cxx::__pool_alloc<wchar_t>::_S_force_new () #7 0x0022f798 in ?? () #8 0x7c96d826 in ntdll!RtlpNtMakeTemporaryKey () from C:\WINDOWS\system32\ntdll.dll #9 0x003e0000 in ?? () #10 0x00448270 in __gnu_cxx::__mt_alloc<wchar_t>::_S_binmap () #11 0x7c96d9dc in ntdll!RtlpNtMakeTemporaryKey () from C:\WINDOWS\system32\ntdll.dll #12 0x003e0000 in ?? () #13 0x00448278 in __gnu_cxx::__pool_alloc<wchar_t>::_S_force_new () #14 0x40000060 in ?? () #15 0x7c96d994 in ntdll!RtlpNtMakeTemporaryKey () from C:\WINDOWS\system32\ntdll.dll #16 0x003e0000 in ?? () #17 0x003e4418 in ?? () #18 0x40000060 in ?? () #19 0x003e0000 in ?? () #20 0x003e4418 in ?? () #21 0x00000000 in ?? () -------------------- Doing a similar build in Visual C++ 2008 with msvc builds of the library, I get the error Unhandled exception at 0x7c812aeb in filesystest.exe: Microsoft C++ exception: boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<boost::filesystem::basic_filesystem_error<boost::filesystem::basic_path<std::basic_string<char,std::char_traits<char>,std::allocator<char> >,boost::filesystem::path_traits> > at memory location 0x0012fb34.. whether I run a debug build or not. -------------------- Compiling with the debug builds of the libraries (X-mgw34-mt-d-1_37.lib) didn't produce significantly-different looking other errors from the ones above. Someone on the internet ( http://forums.codeblocks.org/index.php/topic,8429.msg64317.html#msg64317 ) seemed to have a similar problem, but his solution isn't satisfactory for me (for the reason given below). Linking to single-threaded builds of boost didn't seem to make a difference either. Any help with this would be really appreciated. Stephen *I say 'appears to run fine' because, in the program that I'm trying to port, in windows all I see is a blank window (it's an SDL application); I can std<<cout directory structures from it though. This is most likely an unrelated problem, but I can't start to debug that until I've gotten this problem out of the way.

2 2