November 2004 - Boost-users - lists.preview.boost.org

Re: Re: Re: Leak when using boost threads under VC .NET using usi ng statically linked CRT
by alansa＠ati-uk.com 29 Nov '04

29 Nov '04

-----Original Message----- From: Alan Saunders Sent: 22 November 2004 17:24 To: 'boost-users(a)lists.boost.org' Subject: >> Hmmm no linker errors or warnings. I've set my main exe to statically >> link. if i use the version of threads that statically links to >> msvcrt71 or dynamically to msvcrt71.dll it doesn't seem to make any >> difference. Our boost libraries where compiled here for msvcrt71 and i >> can confirm this. >> >> As i say if my main exe uses msvcrt71.dll then theres no leak. >Are you using tls code? >TLS currently (as to my knowledge) contains a "intentional" leak which >however is harmless. >Could you post the messages? >Did you track down the leak? Hi Roland In order to clear random influences from other areas of the software i'm working on, I have created a simple program with the code that i listed before. nothing else, no tls code. The intentinal leak would be harmless if it did not keep increasing. If I let the program run for long enough i will run out of memory. Sorry which messages would you like me to post? Haven't tracked down the leak yet. My problem is solved for the time being by using msvcrt71.dll as i said before. However it would be nice to be able to using the static lib. I replace the code in the loop which constantly starts and stop threads with my own code that did the same thing. This code was pretty much the same as the code for boost::thread using the windows api functions. This did not leak. However the big difference between my code and boost::thread is that my code didn't use boost::function. Therefore I suspect the leak could be in there! Looking at function it certainly isn't as trivial as thread. One thing i have noticed is that there is an option to get function to use "new" instead of std::allocator with the macro BOOST_NO_STD_ALLOCATOR. (function_template.hpp). I haven't got round to trying this yet though. Alan

4 4

accessing internal graph properties via vertex-iterators directly?
by Sebastian Weber 29 Nov '04

29 Nov '04

Hello there! First of all thanks a lot for the rapid response of my last question. I got another question concerning the access of internal graph properties: To access a bundeled vertice internal properties of a graph one has to use the following syntax: struct VertStruct { int stuff; }; typedef adjacency_list<vecS, vecS, undirectedS, VertStruct> graph_t; graph_t bg; ... setup graph ... graph_traits<graph_t>::vertex_iterator vi, vend; tie(vi, vend) = vertices(bg); bg[*vi].stuff = 4; This makes it clear, that even though we already have an iterator (vi) pointing to our vertice of interest, the BGL has to spend again some time locating the data ascociated with it. A far more direct and much clearer notion would be: (*vi).stuff = 4; With this, one would save a lookup in the data-structures and the vertice-iterator concept would be far more logic to me, concernig internal property handling. I don't know the BGL internals and how this behaviour is optimally programmed with the BGL. I thought of providing a custom container for the vertex-container which is composed of a vector of a customary struct, which glues vertice properties together with every element of the vertice list and makes the property really internal, allowing a (*vi).myrealinternalstuff = 4 expression. I hope my point is clear. Thanks a lot for any help. Greetings, Sebastian Weber

2 2

Graph library - bundled properties for edge
by TC MA 29 Nov '04

29 Nov '04

I try to add Vertex and Edge data with bundled properties. I am modifying the sample program interior_property_map.cpp to learn how to access the bundled properties, but I don't know how to fix my compile error. What is the right way to access the bundled properties for edge? Source code is: #include <boost/config.hpp> #include <iostream> #include <algorithm> #include <boost/graph/adjacency_list.hpp> #include <boost/property_map.hpp> using namespace std; using namespace boost; struct VertexData { int x; int y; }; struct EdgeData { int weight; }; template <class EdgeIter, class Graph> void who_owes_who(EdgeIter first, EdgeIter last, const Graph& G) { // Access the propety acessor type for this graph while (first != last) { // fail to compile here!!! cout << G[source(*first, G)].x << "," << G[source(*first, G)].y << " joins " << G[target(*first, G)].x << "," << G[target(*first, G)].y << " weight " << G[edge(*first)].weight << endl; ++first; } } int main() { // Create the graph, and specify that we will use store the data. typedef adjacency_list<vecS, vecS, directedS, VertexData, EdgeData> MyGraphType; typedef pair<int, int> Pair; // from <utility> Pair edge_array[11] = { Pair(0,1), Pair(0,2), Pair(0,3), Pair(0,4), Pair(2,0), Pair(3,0), Pair(2,4), Pair(3,1), Pair(3,4), Pair(4,0), Pair(4,1) }; MyGraphType G(5); for (int i=0; i<11; ++i) add_edge(edge_array[i].first, edge_array[i].second, G); G[0].x = 0; G[0].y = 0; G[1].x = 1; G[1].y = 1; G[2].x = 2; G[2].y = 2; G[3].x = 3; G[3].y = 3; G[4].x = 4; G[4].y = 4; // store weights in edges graph_traits <MyGraphType>::edge_iterator ei, eend; int i = 0; for (tie(ei, eend) = edges(G); ei != eend; ++ei) { G[*ei].weight = i++; cout << "G[e] weight " << G[*ei].weight << endl; } who_owes_who(edges(G).first, edges(G).second, G); cout << endl; return 0; } Compile error is: gcc-C++-action /home/tcma/cpp/boosttcma/libs/graph/interior_property_map.test/gcc/debug/inlining-on/interior_property_map.o /home/tcma/cpp/boosttcma/libs/graph/interior_property_map.cpp: In function `void who_owes_who(EdgeIter, EdgeIter, const Graph&) [with EdgeIter = boost::detail::adj_list_edge_iterator<boost::counting_iterator<size_t, boost::use_default, boost::use_default>, boost::detail::out_edge_iter<__gnu_cxx::__normal_iterator<boost::detail::sep_<size_t, boost::property<boost::edge_bundle_t, EdgeData, boost::no_property> >*, std::vector<boost::detail::sep_<size_t, boost::property<boost::edge_bundle_t, EdgeData, boost::no_property> >, std::allocator<boost::detail::sep_<size_t, boost::property<boost::edge_bundle_t, EdgeData, boost::no_property> > > > >, size_t, boost::detail::edge_desc_impl<boost::directed_tag, size_t>, ptrdiff_t>, boost::adjacency_list<boost::vecS, boost::vecS, boost::directedS, VertexData, EdgeData, boost::no_property, boost::listS> >, Graph = main()::MyGraphType]': /home/tcma/cpp/boosttcma/libs/graph/interior_property_map.cpp:113: instantiated from here /home/tcma/cpp/boosttcma/libs/graph/interior_property_map.cpp:64: error: no matching function for call to `edge(boost::detail::edge_desc_impl<boost::directed_tag, size_t>)' set -e "g++" -c -Wall -ftemplate-depth-255 -g -O0 -Wno-inline -I"../../../bin/boost/libs/graph/example" -I "/home/tcma/cpp/boost_1_32_0" -o "/home/tcma/cpp/boosttcma/libs/graph/interior_property_map.test/gcc/debug/inlining-on/interior_property_map.o" "/home/tcma/cpp/boosttcma/libs/graph/interior_property_map.cpp" "/usr/bin/objcopy" --set-section-flags .debug_str=contents,debug "/home/tcma/cpp/boosttcma/libs/graph/interior_property_map.test/gcc/debug/inlining-on/interior_property_map.o" ... ===== TingChong Ma ______________________________________________________________________ Post your free ad now! http://personals.yahoo.ca

2 1

RE: Problem with graph:johnson_all_pairs_shortest_paths
by keith.briggs＠bt.com 29 Nov '04

29 Nov '04

>So, you get a code like: const int V = n; int D[V][V]; In this case, 'V' is not a compile time constant and the second line >is not valid -- array bounds must be compile time constant. I have no idea why you get the error you get, instead of >error right at the point where array is declared. Most likely, this is diagnostic problem in the rather ancient (;-)) version of >gcc that you're using. Ok, so what's the right say to do this kind of dynamically-sized problem? Keith

2 1

Problem with graph:johnson_all_pairs_shortest_paths
by keith.briggs＠bt.com 29 Nov '04

29 Nov '04

I am using g++ 3.2 on a redhat linux system. I took the johnson_all_pairs_shortest_paths example, (http://www.boost.org/libs/graph/example/johnson-eg.cpp) which works fine, but then changed const int V = 5; to const int V = n; where n has been computed at runtime. This should not change the type of V or D, but I now get the error below. Can anyone explain how to fix this? Thanks, Keith > /usr/bin/g++ -I ~/boost_1_32_0/ graph_stats.cc graph_stats.cc: In function `int main()': graph_stats.cc:46: no matching function for call to ` johnson_all_pairs_shortest_paths(main()::Graph&, int[V][V], boost::bgl_named_params<int*, boost::vertex_distance_t, boost::no_property> )'

2 1

Serialization and source code organization
by Malte Clasen 28 Nov '04

28 Nov '04

Hi, I'm currently trying to use the new Boost 1.32 serialization library to save the state of a medium complex program (about 10.000 lines of code). The problem is that the serialization lib seems to require very close source code coupling: The translation unit that contains the load/save functions for the application has to see all serialize() method definitions (as they are templates). That means that I cannot rely on forward declarations, so I always have to include the respective header file. This slows down the compilation noticably and results in errors such as "C1204: compiler limit : internal structure overflow" and "C1055: compiler limit : out of keys" (VC++ 7.1), as more or less all significant headers are included in this translation unit. The documentation is aware of this kind of problem and recommends manual instantiations of the serialize() methods (in the PIMPL section). Do I get something wrong or is this the only solution? I don't really like manual template instantiations, but in this case I can't imagine another way to deal with it, except waiting for the next VC++ release and hoping for the best. Malte

4 8

help needed on random libary
by Deng Kun 28 Nov '04

28 Nov '04

can someone tell me how to seed the random library? I am having some problem with the following code the seed method won't compile on solaris but works fine on windows especially in the "else" part die->engine().seed((unsigned int)(s)); gives me some error. #ifndef RANDOMBITSET_H #define RANDOMBITSET_H #include <stdlib.h> #include <boost/random.hpp> #include <boost/dynamic_bitset.hpp> class RandomBitset { public: typedef boost::mt19937 EngineType; typedef boost::bernoulli_distribution<> DistType; typedef boost::variate_generator<EngineType, DistType> DieType; EngineType rng; DistType dist; DieType *die; RandomBitset(double p=0.5): rng(), dist(p){ die= new DieType(rng, dist); } RandomBitset(const RandomBitset & r){ rng=r.rng; dist=r.dist; if (die!=NULL) { delete die; } die= new DieType(rng,dist); } virtual ~RandomBitset(){ delete die; } //force reseed //input s<0 will use system time to reseed //otherwise use s as seed void seed(int s=-1){ if (s<0) { long rs=time(NULL); srand(rs); die->engine().seed((unsigned int) rand()); }else { die->engine().seed((unsigned int)(s)); } } void randomBitset(boost::dynamic_bitset<> & bs); }; #endif //RANDOMBITSET_H

1 0

boost and g++-4.0
by Marc Cromme 28 Nov '04

28 Nov '04

I was wondering if Boost does work with g++ 4.0 (on Linux/Unix systems). I have seen that the config files boost/config/compiler/gcc.hpp check up to version 3.4. (boost version 1.31.0-9 on a debian testing system) Does anybody have experiences with g++ 4.0 ?? Is there a gcc.hpp including tests for g++ 4.0 ?? Something else I should know and do not ?? Marc Cromme -- Marc Cromme, cand. polyt, Ph.D Senior Developer, Project Manager Index Data Aps Købmagergade 43, 2 1150 Copenhagen K. Denmark tel: +45 3341 1000 fax: +45 3341 0101 http://www.indexdata.com INDEX DATA Means Business for Open Source and Open Standards

1 0

Can't compile very simple snippet that uses boost::tuple
by Frank Astier 28 Nov '04

28 Nov '04

I am pretty sure it's because of my lack of skills with templates... Still, I don't understand why I can stream a vector if it's on its own, but not if it's stored in a boost::tuple. And it works with a std::pair! Any hint? #include <boost/tuple/tuple.hpp> #include <boost/tuple/tuple_io.hpp> #include <iostream> #include <vector> #include <algorithm> using namespace std; using namespace boost; template <typename T> ostream& operator<<(ostream& out, const vector<T>& v) { copy(v.begin(), v.end(), ostream_iterator<T>(out, " ")); return out; } template <typename T, typename U> ostream& operator<<(ostream& out, const pair<T, U>& p) { return out << p.first << " " << p.second; } int main(int, char**) { vector<float> v; cout << v; // Works fine pair<vector<float>, pair<int, int> > p; cout << p; // Works fine tuple<vector<float> > t; cout << t; // <== WON'T COMPILE } Thanks, Frank

2 1

Another use for apply.
by Paul Elliott 28 Nov '04

28 Nov '04

The bind documentation mentions the apply object as a way to get around the fact that bind does not evaluate its first arguement. There is another use. You can reverse the roles of function and parameters. You can use apply and bind to turn a parameter into a functor, and call it using a function as a parameter! Example. #include <boost/bind.hpp> #include <boost/function.hpp> #include <boost/bind/apply.hpp> #include <fstream> #include <iostream> #include <iomanip> // some sample test functions. (data) int g(int x) {return x+1;}; int h(int x) {return x*x; }; int main(int argc, char * argv) { // want to work with functions of the form int f(int); typedef int P(int); // bind the integer 5 to apply turning // apply into the number 5 into a function! call it five boost::function<int (P)> five = boost::bind(boost::apply<int>(),_1,5); //int y = five(g); // Now call 5 with the "data" g and h. std::cout << "+1=" << five(g) << std::endl << "square = " << five(h) << std::endl; return 0; }; Results: +1=6 square = 25 -- Paul Elliott 1(512)837-1096 pelliott(a)io.com PMB 181, 11900 Metric Blvd Suite J http://www.io.com/~pelliott/pme/ Austin TX 78758-3117

1 0