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

Compilation problems on Linux
by mayankg＠iiitb.ac.in 31 Mar '04

31 Mar '04

Hi, We are trying to use the libboost_python.so library in Linux with our project. When we try to compile the C++ file of our project having all the boost interfaces, we get the following compilation error only on Linux. Although on Windows it works fine. We are using the Python 2.2.3 and Boost 1.30.2 versions. /usr/include/boost/python/converter/as_to_python_function.hpp:28: sorry, unimplemented: ` method_call_expr' not supported by dump_expr /usr/include/boost/python/converter/as_to_python_function.hpp:28: sorry, unimplemented: ` method_call_expr' not supported by dump_expr /usr/include/boost/python/converter/as_to_python_function.hpp:28: sorry, unimplemented: ` method_call_expr' not supported by dump_expr /usr/include/boost/python/converter/as_to_python_function.hpp:28: sorry, unimplemented: ` method_call_expr' not supported by dump_expr Please help, Regards, Mayank

2 1

Location and namespace of a boost library (say, iterators)
by Deane Yang 31 Mar '04

31 Mar '04

I'm sorry if I'm being dumb about this, but I've run into the following difficulty a couple of times with boost libraries. I've decided to try using the new boost iterator library. Based on the documentation, I think I want to use iterator_adaptor. But scanning through the documentation, I don't immediately see - what namespace iterator_adaptor lives in. I assume it's just boost, but it probably doesn't hurt to mention this explicitly, since some boost libraries use sub-namespaces. - what header file to include. I couldn't find this in the documentation. I can figure all of this out, but it is helpful to have it specified in the documentation.

2 1

Installing Boost 1_30_0 on HPUX
by Albert Gome 31 Mar '04

31 Mar '04

Hi, I am new to BOOST and i like to install that on HPUX11.11. I have downloaded BOOST_1_30_0 and have gcc installed on HPUX. What is the efficient way of installing it. ie the parameters with the bjam. Thanks Albert __________________________________ Do you Yahoo!? Yahoo! Finance Tax Center - File online. File on time. http://taxes.yahoo.com/filing.html

2 1

[BGL] Finding out what properties exist in a graph?
by Steve Juranich 30 Mar '04

30 Mar '04

Is there an easy way to find out all of, for instance, the vertex properties associated with a graph? What I'm looking for is something like the following: <pseduocode> vector<properties> my_vertex_properties; property_iterator pi, last; for(tie(pi, last) = vertex_property_list(Graph); pi != last; ++pi) my_vertex_properties.push_back(*pi); </pseudocode> How about it? Does anything like this exist or am I just daydreaming again? Thanks. -- Stephen W. Juranich Science Applications Intl. Corp. (SAIC) (520) 570-7706

2 2

Serialization library
by Steve Folly 30 Mar '04

30 Mar '04

Hi, I'm looking using the Robert Ramey's serialization library for our own projects and would like to know it's current status. Obviously, it's not in 1.31.0, so what are the plans for it's future inclusion? Thanks. Steve.

2 1

graph, dfs search
by Peter Aronsson 30 Mar '04

30 Mar '04

Hi. I am having a problem with depth_first_search. The thing I do different is that I do not put any info in a property map and I do not want to bother myself with colormaps etc. The application code does: Graph g; pass2_visitor vis; depth_first_search(g,boost::visitor(vis)); But I get a compilation error, saying that it can not find a suitable operator+ in the property_map... See attached files for complete example. Grateful for all help! (I am running from HEAD on sourceforge cvs ) Regards Peter Aronsson -- _________________________________________________________________ / Peter Aronsson, Phd Student at PELAB (Programming Environments \ | Laboratory ) Department for Computer & Information Science | | Linköping University, Sweden | |=================================================================| | petar(a)ida.liu.se , phone +46 (0)13-28 1737 Room 3B:490 | \_________________________________________________________________/ #include "tg.hpp" int main() { TaskGraph g; pass2_visitor vis; depth_first_search(g,boost::visitor(vis)); } #ifndef _TASKGRAPH_H #define _TASKGRAPH_H #include <boost/config.hpp> #include <iostream> // for std::cout #include <utility> // for std::pair #include <algorithm> // for std::for_each #include <boost/utility.hpp> // for boost::tie #include <boost/graph/graph_traits.hpp> // for boost::graph_traits #include <boost/graph/adjacency_list.hpp> #include <boost/graph/adjacency_iterator.hpp> #include <boost/graph/graphviz.hpp> #include <boost/graph/depth_first_search.hpp> //Task graph types using namespace std; using boost::tie; struct vertex_execcost_t { typedef boost::vertex_property_tag kind; }; struct vertex_unique_id_t { typedef boost::vertex_property_tag kind; }; struct edge_result_set_t { typedef boost::edge_property_tag kind; }; typedef boost::property<boost::vertex_name_t,string, boost::property<vertex_execcost_t,float, boost::property<vertex_unique_id_t, int > > > VertexProperty; typedef boost::property<boost::edge_weight_t, int > EdgeProperty; typedef boost::adjacency_list<boost::listS, boost::listS, boost::bidirectionalS, VertexProperty, EdgeProperty> TaskGraph; typedef boost::graph_traits<TaskGraph>::vertex_descriptor VertexID; class pass2_visitor : public boost::default_dfs_visitor { public: pass2_visitor(/* initialize containers to store stuff in*/) { }; template <typename Vertex, typename Graph> void discover_vertex(Vertex u, const Graph & g) const { // Do some stuff }; }; #endif

2 4

CVS import gotcha - beware!
by Steve Folly 29 Mar '04

29 Mar '04

Hi, Most of you are probably aware of this, but I just thought I'd share my experience with trying to import the new version of boost into our CVS repository using 'cvs import': It wasn't recognising new folders - e.g. boost/spirit/core and all it's contents. I was puzzling it for a while until I realised the import output was showing 'I' next to boost/spirit/core. Aha - it's ignored. DOH! Well, core files are ignored, aren't they! To all those who have been long suffering with this : don't forget to include '-I !' in the command to include all files!! I'll be happy if it helps at least one person! :) -- Regards, Steve.

1 0

RE: [Boost-users] Defining less on member variables/functions
by Chris Jurney 29 Mar '04

29 Mar '04

In this case, I could have used pair's operator<, but if I have a class by value as the second item of the pair that doesn't have a meaningful operator<, I want to ignore it in the sort. Thanks for all the help, guys. Chris -----Original Message----- From: boost-users-bounces(a)lists.boost.org [mailto:boost-users-bounces@lists.boost.org] On Behalf Of Victor A. Wagner Jr. Sent: Friday, March 26, 2004 8:40 AM To: boost-users(a)lists.boost.org Subject: Re: [Boost-users] Defining less on member variables/functions why do you want to sort ONLY on the 1st element of the pair? At Thursday 2004-03-25 11:34, you wrote: >content-class: urn:content-classes:message >Content-Type: multipart/alternative; > boundary="----_=_NextPart_001_01C41297.D800D40A" > >I've had this same pattern come up 4 or 5 times now, and each time I've >been unable to find a stl + boost solution that allowed me to do what I >want without writing my own less function. Basically, I want something >that will let me do a sort, or define a map on structs or pairs without >writing a less that compares the member variable or ..first. My last attempt: > >vector<pair<float, float>> myVector >std::sort(myVector.begin(), myVector.end(), _1.first < _2.first); > >No such luck. It looks like I can do operator * and & on lambda objects, >but not access members through them. Is there anything in boost, stl, or >elsewhere that lets me not write this: > >class myLess >{ > bool operator() (pair<float, float> a, pair<float, float>b) > {return a.first < b.first;} >} > >Thanks for any help, >Chris > > >_______________________________________________ >Boost-users mailing list >Boost-users(a)lists.boost.org >http://lists.boost.org/mailman/listinfo.cgi/boost-users Victor A. Wagner Jr. http://rudbek.com The five most dangerous words in the English language: "There oughta be a law" _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

Microsoft Managed C++
by Roger Wilco 29 Mar '04

29 Mar '04

I have a somewhat large (~100K lines) C++ server application. It uses Boost (lots of smart pointer use, thread use, regex), Xerces XML libraries, some MFC for GUI, and ADO for database access. I tried to build a test build of my application using Microsoft's Managed C++. I got all my code to compile, but I can't link to either Boost or Xerces XML. Is this possible? Has anyone done this? If I ultimately go the .NET route, I would replace Boost and Xerces with functionality provided by the native .NET runtime libraries. However, that would be a massive time consuming undertaking. At this point, I just want to dip my toes in the water and observe pros/cons of .NET, and have the option of migrating to .NET piecemeal. I have solid experience with C# and Java and I am extremely experienced with C++ (10 years). My main problems with C++ are: - Compile times. On my 2.8GHz dev system, a rebuild takes ~20 minutes and even a build after making a minor change takes 45 seconds. It's amazing how VC++ build speeds really have remained so stagnant despite increases in hardware speed. I realize VC++ is the slower of the IDEs for build times. - Debugging memory corruption or thread concurrency issues. This is much easier in Java/C# and the tools for assisting are much easier to use. I spent many days trying tools such as Purify for C++ and couldn't produce any fruitful results. - Diagnosing memory usage. My server application uses lots of memory. After running for several hours under production conditions it will reach around 400MB-1.3GB depending on the tasks its doing. I'd like to get a general overview of where this memory is being used. I don't suspect blatant memory leaks (new with no delete) but instead runaway caches or legitimate usage. These issues are of course still present in C#/Java but the diagnostic tools to address these issues seem to be much better. I haven't found any usable C++ tool for this issue and I rely on manually logging data points and manually trying to infer an idea of what is happening. - Java/C# is overall a better (easier/faster to develop/maintain) development platform for this type of server application work. However, I will happily stick with C++ if I can resolve these other issues. And I am of course, concerned about any performance hit when migrating to Java/C#. We max out CPU and RAM use, so these factors aren't important. I couldn't take more than a 25% drop in efficiency. Any other suggestions on my situation?

2 1

SQLite
by Witz 28 Mar '04

28 Mar '04

I don't know if this is the forum for bringing this up but ... I was wondering if anyone thinks SQLite might make a good addition to Boost - www.sqlite.org - if the author would agree to it. Ian Whittley

2 1

boost::format with non-standard string class
by Scott Lamb 27 Mar '04

27 Mar '04

Hi, I'm not real happy with std::string, largely because it seems poorly suited to UTF-8. I'm happier with other classes such as Glib::ustring. [1,2] Is it possible to use boost::format with such a class? I'd previously thought that boost as a whole expected you to use std::string, but the recent discussion of boost::regex + Glib::ustring gave me some hope. It seems the requirements for boost::regex are more simple - a bidirectional, read-only character (not byte) iterator. That should be easy enough to provide for any UTF-8 string class. I do see explicit mention of std::basic_string in the boost::format headers, but I thought I'd ask anyway. (I likely will find or make another class to avoid pulling in all of glib. But it would be similar to Glib::ustring.) [1] - more in-depth discussion of Glib::ustring vs. std::string at <http://www.gtkmm.org/gtkmm2/docs/reference/html/ classGlib_1_1ustring.html#_details> [2] - additional reasons for disliking std::string at <http://www.sgi.com/tech/stl/string_discussion.html> Thanks, Scott

1 0

boost/graph/push_relabel_max_flow::is_flow() bug?
by Arun Bhalla 27 Mar '04

27 Mar '04

Hi, I'm not really a user of Boost, but I've been referring to the Boost Graph Library code for another project. In particular, I've been looking at push_relabel_max_flow.hpp. I was writing a unit test which followed the same ideas of the is_flow() test, and it was failing on what should be a flow network with a valid maximum flow. The code in question: // a is an edge_descriptor if (capacity[a] > 0) if ((residual_capacity[a] + residual_capacity[reverse_edge[a]] != capacity[a]) || (residual_capacity[a] < 0) || (residual_capacity[reverse_edge[a]] < 0)) return false; I contend the test should actually be: // a is an edge_descriptor if (capacity[a] > 0) if ((residual_capacity[a] + residual_capacity[reverse_edge[a]] ! != capacity[a] + capacity[reverse_edge[a]]) || (residual_capacity[a] < 0) || (residual_capacity[reverse_edge[a]] < 0)) return false; I refer to an example network in CLR (1st ed.), p.589. There are several vertices, but I focus on vertices v1 and v2. In this graph, we have capacities c(v1,v2) = 10 c(v2,v1) = 4 and residual capacities cf(v1,v2) = 11 cf(v2,v1) = 3 With the above (original) test, it would fail because 11 + 3 (the sum of residual capacities) != the capacity of either edge. This would hold true if the sum of residual capacities were compared with the sum of capacities. Ordinarily this test would work if reverse_edge[a] has a capacity of 0, which would be the case if the edge (a) is in the flow network, but its reverse is not. I'm not a graph theory expert, though, so I defer to anyone who is. :) Thanks, Arun

1 0

Defining less on member variables/functions
by Chris Jurney 26 Mar '04

26 Mar '04

I've had this same pattern come up 4 or 5 times now, and each time I've been unable to find a stl + boost solution that allowed me to do what I want without writing my own less function. Basically, I want something that will let me do a sort, or define a map on structs or pairs without writing a less that compares the member variable or .first. My last attempt: vector<pair<float, float>> myVector std::sort(myVector.begin(), myVector.end(), _1.first < _2.first); No such luck. It looks like I can do operator * and & on lambda objects, but not access members through them. Is there anything in boost, stl, or elsewhere that lets me not write this: class myLess { bool operator() (pair<float, float> a, pair<float, float>b) {return a.first < b.first;} } Thanks for any help, Chris

6 8

Re: [Boost-users] MAC OS X installation of boost libs
by Michael Short 26 Mar '04

26 Mar '04

Hello I am using MAC OS X 10.3.3. First I done find the bjam tool? Second when I build the bjam tool myself I get the following error. Jamfile:71: in module scope rule unless unknown in module. I think if you could please just give me a quick step by step of what you did to setup the libs with the correct directorys you used then that would probably fix the problems. I did try following the instructions on the page you stated but just did not help. thanks Michael On Tuesday, March 23, 2004, at 09:25PM, James Mastro <jmastro(a)rochester.rr.com> wrote: >On Mar 22, 2004, at 5:12 PM, Michael Short wrote: > >> The only problem is that I am having problems installing and building >> the libs on my MAC OS X system. > >What kind of problems? What version of OS X? I built boost 1.31.0 on >OS X 10.3.2. I think the regex library gave me some problems, but I >believe everything else built OK. > >I just followed the directions here: > >http://www.boost.org/more/getting_started.html > >-jim > > >

4 7

MultiArray questions
by Dr Mark H Phillips 24 Mar '04

24 Mar '04

Hi, I've recently found out about the existence of MultiArray and I am quite impressed! I was thinking of implementing something similar myself but now hopefully I don't have to! However I have a few questions that I'm hoping someone can answer. Question 1 ---------- In the user documentation there is an example which has some example code: for(index i = 0; i != 3; ++i) for(index j = 0; j != 4; ++j) for(index k = 0; k != 2; ++k) A[i][j][k] = values++; Why can't you change the != into <? Ie, do: for(index i = 0; i < 3; ++i) for(index j = 0; j < 4; ++j) for(index k = 0; k < 2; ++k) A[i][j][k] = values++; or can you? This would be more efficient I should think, and requiring that index types be ordered shouldn't be too onerous should it? Question 2 ---------- If I use the built-in "no-frills" C array type to represent a 2-dimensional array in row-major order, then I would access an element (i, j) by doing: a[ &a[0] + i*n + j ] where a is my m by n array. And if I wish to traverse a, in row-major order, I can do: for (int* ip=&(a[0]); ip<&(a[n*m]); ip+=n) { for (int* ijp=&(ip[0]); ijp<&(ip[n]); ++ijp) { visit(ijp); } } These are a very efficient ways of doing things, but it lacks somewhat in ease-of-generalization and in readbility. This is where the MultiArray library is useful. But do I lose in efficiency by going to MultiArray? Ie, in the simple case where base is 0, stride is 1, and when we traverse in row-major order, is the MultiArray library smart enough to do compile-time optimizations to generate code which is as efficient as the above? I am guessing that this sort of approach is what the Blitz++ library does?? Question 3 ---------- Is it possible to construct a new multi_array/multi_array_ref/const_multi_array_ref from an array_view by doing a deep copy of the view? A view is a nice way of looking at different parts of a MultiArray object in a readable way, but sometimes you might wish to turn this view into a more-efficiently represented object. Is this easy to do in the MultiArray library? Question 4 ---------- Sometimes you might have an object starts off being mostly accessed in row-major order, so you want a representation which works efficiently for this use. But down the track you might wish to change to mostly column-major access. Does MultiArray provide easy ways to transform from one underlying representation to another? Thanks, Mark Phillips.

3 2

RE: [Boost-users] boost::filesystem::exists() & GetfileAttributesA
by Ben Hutchings 24 Mar '04

24 Mar '04

Sebastian Orfino <sorfino(a)lanacion.com.ar> wrote: > I am experiencig a problem with exists() and its Windows > implementation with GetFileAttributes API. I am using exists() prior > to the scan of a directory to see if it exists; the fact is that on > W2K (just tested it on that OS) after a day or two > GetFileAttributesA() returns INVALID_FILE_ATTRIBUTES but the path > does exists (it is a UNC-style path). > After a restart of the app, all works fine. <snip> It couldn't possibly be related to this bug, could it: <http://support.microsoft.com/default.aspx?scid=kb;en-us;325929>? I know that that talks about Windows XP clients and NT 4 servers but MS does not always test thoroughly whether a bug found in one version of a product exists in other versions. The bug is supposed to be fixed in Windows XP SP2, by the way.

1 0

It always show compiler limitation when I use boost.python to wrap a class under MSVisualStudio .Net
by Shunming Fang 24 Mar '04

24 Mar '04

Hello, guys, I try build my c++ lib as python modules using boost.python, when compile one big class, it always show compiler limit( my ram already update to 1.5 GB) while the class can be smoothly build on other platforms like SGI, Linux(RAM 512 MB+). I don't know why? I already upgrade my MSVS .net to 2003Version, Anybody know How I can get around this problems? Thank you very much __________________________________ Do you Yahoo!? Yahoo! Finance Tax Center - File online. File on time. http://taxes.yahoo.com/filing.html

1 0

Boost error on Red Hat Linux 7.4
by Jonathan Arnold 24 Mar '04

24 Mar '04

I'm trying to build the latest (1.31.0) Boost libraries for the first time on a Red Hat Linux 7.4, gcc 2.96 system. It seems to build all the libs and such correctly, but I get the following error trying to build the test programs: gcc-C++-action bin/boost/libs/test/build/libboost_test_exec_monitor.so/gcc/debug/shared-linkable-true/test_tools.o /home/jdarnold/boost_1_31_0/libs/test/build/../src/test_tools.cpp: In method `void boost::test_toolbox::detail::print_log_value<char>::operator() (ostream &, char)': /home/jdarnold/boost_1_31_0/libs/test/build/../src/test_tools.cpp:194: parse error before `__ctype_b' /home/jdarnold/boost_1_31_0/libs/test/build/../src/test_tools.cpp:205: confused by earlier errors, bailing out set -e g++ -c -Wall -ftemplate-depth-100 -DBOOST_TEST_NO_AUTO_LINK=1 -g -O0 -fno-inline -fPIC -I"bin/boost/libs/test/build" -I "/home/jdarnold/boost_1_31_0" -o "bin/boost/libs/test/build/libboost_test_exec_monitor.so/gcc/debug/shared-linkable-true/test_tools.o" "/home/jdarnold/boost_1_31_0/libs/test/build/../src/test_tools.cpp" /usr/bin/objcopy --set-section-flags .debug_str=contents,debug "bin/boost/libs/test/build/libboost_test_exec_monitor.so/gcc/debug/shared-linkable-true/test_tools.o" What is going wrong here? Is my version of F++ too old? Thanks in advance. -- Jonathan Arnold (mailto:jdarnold@buddydog.org) Daemon Dancing in the Dark, a FreeBSD weblog: http://freebsd.amazingdev.com/blog/

1 0

Cannot use BGL & boost::format together under VS2003
by Alexeev Gleb 24 Mar '04

24 Mar '04

Hi all! The following simple sample doesn't compile on VC++ .NET 2003: #include <iostream> #include <boost/format.hpp> // without the following line everything's ok #include <boost/graph/adjacency_list.hpp> int main(){ std::cout << boost::format("%d, %s") % 5 % "some text"; return 0; } -- Best regards, Alexeev Gleb mailto:alexeev_gleb@hotbox.ru

2 2

boost/graph - induced subgraph
by Matthew Galati 24 Mar '04

24 Mar '04

Is it possible to find the minimum spanning tree of an induced subgraph? vector< graph_traits<Graph>::vertex_descriptor > vd; typedef subgraph< Graph > SubGraph; SubGraph & inducedG = subG.create_subgraph(vd.begin(), vd.end()); Here is the example I am looking at: induced verts: 1 2 3 4 5 7 Original Graph: 0 <--> 1 2 5 7 1 <--> 0 3 5 7 2 <--> 0 3 4 5 3 <--> 2 1 4 5 4 <--> 2 3 5 6 5 <--> 0 1 2 3 4 6 7 6 <--> 4 5 7 7 <--> 0 1 5 6 Induced Graph: 1 <--> 3 5 7 2 <--> 3 4 5 3 <--> 2 1 4 5 4 <--> 2 3 5 5 <--> 1 2 3 4 7 7 <--> 1 5 So, the "create_subgraph" seems to be working fine. But, when I run kruskals with: vector<graph_traits < SubGraph >::edge_descriptor> spanning_tree; kruskal_minimum_spanning_tree(inducedG, back_inserter(spanning_tree)); for (vei = spanning_tree.begin(); vei != spanning_tree.end(); ++vei) { pair<int,int> ij = incident(*vei, inducedG); int edge_index = e_index[*vei]; cout << "e_index: " << edge_index << " : " << ij.first << " " << ij.second << endl; I get a spanning tree of just 4 edges (while the subgraph has 6 vertices). Also, how does the edge_index of the subgraph correspond to the original graph? For example, edge (4,1) is not in the original graph, but it shows up in my spanning tree. Spanning Tree: e_index: 3 : 2 0 e_index: 7 : 4 0 e_index: 4 : 3 1 e_index: 8 : 4 1 Thanks, Matthew -- Matthew Galati ISE Lehigh University IBM Service Parts Solutions 610.758.4042 (Office) 610.882.0779 (Home) magh(a)lehigh.edu, magal11(a)us.ibm.com http://sagan.ie.lehigh.edu/mgalati/

4 3

boost-sandbox socket and boost 1.31.0?
by Robert Helgesson 23 Mar '04

23 Mar '04

Howdy, I'm wondering regarding the socket library found in boost-sandbox. Will it be updated to function nicely with the latest Boost release? The problem seems to be that boost::default_iterator_policies and boost::iterator_adaptor no longer work like the socket library expect. Regards -- / Robert Helgesson |-- rycee(a)home.se --\ \-- GPG Key ID: 41E95A5B --| http://www.rycee.cx/ /

3 2

Re: [Boost-users] obtaining vertex descriptor from property_map (newbie)
by Gustavo Sánchez 23 Mar '04

23 Mar '04

I tend to use a map<string, vertex_descriptor> for this. Doug Thanks for the idea, I was just thinking something like this, but I have a problem. I need the definition of the Graph typedef property<edge_name_t, std::string> EdgeProperty; typedef property<vertex_name_t, std::map<string,vertex_property??> > NameProperty; typedef property<vertex_index_t, std::size_t, NameProperty> VertexProperty; typedef adjacency_list<vecS, vecS, undirectedS, VertexProperty, EdgeProperty> UGraph; in order to get the vertex_descriptor, but I need the map<string,vertex_descriptor> for the VertexPorperty so I drop in a circular type dependency. Can you please be more verbose? thanks --------------------------------- Antivirus Filtros antispam 6 MB gratis ¿Todavía no tienes un correo inteligente?

2 1

Solaris and thread_resource_error
by Peter Jones 23 Mar '04

23 Mar '04

On Feb 1 of this year Baptiste Lepilleur posted a status report for Boost 1.31.0 Release candidate 2 on Solaris and AIX. It was mentioned that "boost.thread is now systematically failing with 'thread resource error'". I am experiencing the same thing on Solaris 8, with GCC 3.3.2 (GNU binutils) and Boost 1.31.0. Has anyone looked into this yet? If not, I am willing to track down what line in Boost.Threads is throwing thread_resource_error(). -- Peter Jones [unix;vi;crypto;c++;4wd] (http://pmade.org) PGP [5137 173E 5BFA 8C63 8982 F035 FDBB 1636 7D69 C63F]

2 3

linking to boost::thread
by Jerry 23 Mar '04

23 Mar '04

Hi, I've just built boost1.31.0 using bjam under Linux, apparently successfully. There don't appear to be any static libraries for boost::thread. Is this correct? If so, could someone please explain why I'm getting linker errors under g++ if I don't mention the _dynamic_ thread libs on the g++ command line, and "cannot open shared object file" at runtime if I do. Any pointers appreciated. TIA, Jerry.

2 1

shared_ptr help needed, please!
by Eric Damphousse 23 Mar '04

23 Mar '04

hi all, quick question concerning shared_ptr... not sure if things get deleted (destroyed) properly and its freaking me out. I hope this is not too long Basically, I use the BGL to store my data and that seems to be ok. The way the BGL stuff is defined is as such... Note that spObjectNode is defined as typedef boost::shared_ptr<CObject> spObjectNode // Set vertex properties typedef boost::property<boost::vertex_name_t, std::string> node_name_t; typedef boost::property<graph_node_t, spObjectNode, node_name_t> node_property_t; // Set edge properties typedef boost::property<boost::edge_name_t, std::string> connection_name_t; typedef boost::property<graph_connection_t, spObjectNode, connection_name_t> edge_property_t; // Set graph typedef adjacency_list<listS, vecS, bidirectionalS, node_property_t, edge_property_t> graph_t; // graph_t m_Graph; The function below is what I use to add a node to the BGL... template <typename T> void AddNode(const T& object) { // Create the vertex boost::shared_ptr<T> pNode(new T(object)); // Add the vertex + its name to the graph vertexdescriptor_t newNode = boost::add_vertex(node_property_t(pNode, node_name_t(pNode->GetObjectName())), m_Graph); }; The way I use the AddNode ... for example in a main.cpp: /// blablabla above CSphere sphere("sphere1"); CSceneDB::Instance().AddNode(sphere); return 0; For testing purposes, in the base class and derived classes (see below) I print a message in the destructor of the class, just to make sure things are ok. If I just have in my main the previous call then upon exit I expect and get the following message... CSceneDB cleanup CSphere cleanup CGeometry cleanup CObject cleanup Now for the source of my anxiety... I wrote a function to get back my object from the BGL and it goes like this... inline boost::shared_ptr<CObject> getNodeProperty(const vertexdescriptor_t& vertex) { boost::property_map<graph_t, graph_node_t>::type node_prop_map; node_prop_map = boost::get(graph_node_t(), m_Graph); return node_prop_map[vertex]; } boost::shared_ptr<CObject> CSceneDB::getNodebyName(const std::string& strName) { boost::shared_ptr<CObject> node_prop; boost::graph_traits<graph_t>::vertex_iterator vi, vi_end; boost::tie(vi, vi_end) = boost::vertices(m_Graph); for(; vi != vi_end; ++vi) { node_prop = getNodeProperty(*vi); if(node_prop != 0) { if(node_prop->GetObjectName() == strName) break; } } return node_prop; } template <typename T> boost::shared_ptr<T> getNode(const std::string& strName) { boost::shared_ptr<CObject> abc = getNodebyName(strName); shared_ptr<T> def = boost::dynamic_pointer_cast<T>(abc); return def; } The problem is that if I do something like this... in the main.cpp /// blablabla stuff above CSphere sphere("sphere1"); CSceneDB::Instance().AddNode(sphere); boost::shared_ptr<CSphere> abc = CSceneDB::Instance().getNode<CSphere>("sphere1"); return 0; I only get upon exit... CSceneDB cleanup what happened to CSphere cleanup, CGeometry cleanup and CObject cleanup, are they leaking? Now for a sanity check, if I do in the main.cpp ... i do NOT do the assignment boost::shared_ptr<CSphere> abc = /// blablabla stuff above CSphere sphere("sphere1"); CSceneDB::Instance().AddNode(sphere); CSceneDB::Instance().getNode<CSphere>("sphere1"); return 0; upon exit i get ... CSceneDB cleanup CSphere cleanup CGeometry cleanup CObject cleanup Can someone tell me why this is, is my assignment wrong? Any help would be greatly appreciated Eric Brief description of the classes used ofr my test // Object.h class CObject { protected: CObject(){}; protected: virtual ~CObject(){cout << "CObject cleanup" << endl;}; }; typedef boost::shared_ptr<CObject> spObjectNode; // Geometry.h class CGeometry : public CObject { protected: CGeometry(){}; protected: virtual ~CGeometry(){cout << "CGeometry cleanup" << endl;}; }; // Sphere.h class CSphere : public CGeometry { public: CSphere(){}; ~CSphere(){cout << "CSphere cleanup" << endl;}; };

3 4