July 2004 - Boost-users - lists.boost.org

using DFS visitors
by Gustavo Sÿffffe1nchez 20 Jul '04

20 Jul '04

hello all, we are trying to use a DFS visitor in a similar way as the BFS visitor which is used in the example "bfs.cpp" (boost 1.31.0), that is the code, but it complains about a lot of errors in the internal boost library files... #include <boost/config.hpp> #include <assert.h> #include <vector> #include <algorithm> #include <utility> #include <iostream> #include <boost/graph/adjacency_list.hpp> #include <boost/graph/depth_first_search.hpp> #include <boost/graph/visitors.hpp> #include <boost/property_map.hpp> #include <boost/graph/graph_utility.hpp> using namespace boost; using namespace std; // visitor que muestra el orden en que se encuentran los arcos template <class NewGraph, class Tag> struct muestra_orden : public boost::base_visitor<muestra_orden<NewGraph, Tag> > { typedef Tag event_filter; muestra_orden(NewGraph& graph) : new_g(graph) { } template <class Edge, class Graph> void operator()(Edge e, Graph& g) { cout << "arco: " << e << endl; } private: NewGraph& new_g; }; template <class NewGraph, class Tag> inline muestra_orden<NewGraph, Tag> MuestraOrden(NewGraph& g, Tag) { return muestra_orden<NewGraph, Tag>(g); } //**************************************************************************+ template <class NewGraph, class Tag> struct muestra_ordenV : public boost::base_visitor<muestra_ordenV<NewGraph, Tag> > { typedef Tag event_filter; muestra_ordenV(NewGraph& graph) : new_g(graph) { } template <class Vertex, class Graph> void operator()(Vertex v, Graph& g) { cout << "ARCO: " << v << endl; } private: NewGraph& new_g; }; template <class NewGraph, class Tag> inline muestra_ordenV<NewGraph, Tag> MuestraOrdenV(NewGraph& g, Tag) { return muestra_ordenV<NewGraph, Tag>(g); } /********************************************************/ int main(int , char* []) { using namespace boost; typedef boost::adjacency_list< boost::mapS, boost::vecS, boost::undirectedS, boost::property<boost::vertex_color_t, boost::default_color_type, boost::property<boost::vertex_degree_t, int, boost::property<boost::vertex_in_degree_t, int, boost::property<boost::vertex_out_degree_t, int> > > > > Graph; Graph G(11); boost::add_edge(5, 6, G); boost::add_edge(0, 1, G); // 0 es el + boost::add_edge(1, 2, G); boost::add_edge(1, 3, G); boost::add_edge(2, 4, G); boost::add_edge(3, 5, G); boost::add_edge(4, 6, G); boost::add_edge(5, 7, G); // 7 es el - boost::add_edge(5, 10, G); boost::add_edge(10, 11, G); boost::add_edge(11, 12, G); typedef graph_traits<Graph>::vertex_descriptor Vertex; typedef graph_traits<Graph>::vertices_size_type size_type; Graph G_copy(11); size_type d[11]; std::vector<Vertex> p(num_vertices(G)); std::fill_n(d,11,0); Vertex s=7; p[s]=s; boost::depth_first_search (G, visitor(boost::make_dfs_visitor( std::make_pair(boost::record_predecessors(&p[0],boost::on_tree_edge()), std::make_pair(MuestraOrdenV(G_copy,boost::on_finish_vertex()), MuestraOrden(G_copy, boost::on_tree_edge())))) ), s); // we can compile the visitor(...) clause alone, but not inside depth_first_search (...) clause. return 0; } thanks a lot Gus --------------------------------- [input] [input] [input]

1 0

Dependency of Boost library on STD namespace
by anas munir 19 Jul '04

19 Jul '04

Dear users, I am trying to port Boost Library with Microtec Toolkit. Is it possible to build Boost Library without Standard C++ RTL placed in std namespace ? If yes, then please guide me to the mechanism to do so. Like in STLport, this thing is configurable. Regards, Anas

2 1

Re: [Boost-users] serialization of a complex object
by Bo Peng 19 Jul '04

19 Jul '04

On Mon, Jul 19, 2004 at 05:08:02AM -0700, Jeff Garland wrote: > There are still some open issues with getting all the serialization tests > working. Thanks. I should have checked this before I used the library.... this cost me several hours of painful debugging. :-( Bo

1 0

serialization of a complex object
by Bo Peng 19 Jul '04

19 Jul '04

Dear list, I have tried many ways but could not figure out how to boost::serialize the following object: Basically, to maintain a big chunk of small objects, I use a memory pool to allocate a big trunk of memory and return the pointers to small objects. struct small { int size; int * ptr; init(int s, int* p){ size = s; ptr = p; } } struct All { vector<int> data; vector<small> objs; All(int num_obj, int size_obj) { data.resize( num_obj * size_obj); for(int i=0; i<num_obj; i++) objs[i].init( size_obj, &data[i* size_obj]); } } It is easy to serialize All::data but how to deal with small::ptr is beyond me. I tried the following ways: 1. in small ar & size & ptr. --- template expansion error. 2. in small ar & size & reinterpret_cast<unsigned int>ptr --- core dump. What I am trying to do in 2 is that in load(....), I want to ar & objs; /// get the number of old pointer back. find the smallest number of objs[i].ptr relocate objs[i].ptr to &data[0] + objs[i].ptr - smallest point. However, I can not do this in the load(...) function. The template expansion does not allow changes to objs (this does not make sense to me though). 3. ignore ptr --- can compile, core dump. Can someone give me a bit of insights how I can make things work? Thanks.

2 3

RE: [Boost-users] Re: shared_ptr: LWM versus atomic_count?
by Ben Hutchings 19 Jul '04

19 Jul '04

Russell Hind <rhind(a)mac.com> wrote: > Ben Hutchings wrote: > > > > It isn't fundamentally necessary to have two counters, though; see > > <http://groups.google.com/groups?threadm=slrnbhbnpc.154.do-not-spam > > -benh%40tin.bwsint.com>. > > I'd seen this (I think you posted the same link recently). I don't > know if anything has been done to change the implementation of > shared_ptr to use this method, though (nothing to do with me). The difficulty is that it depends on a couple of unportable atomic functions that can't be implemented using the existing contents of Boost.Threads or pthreads (though they do translate neatly into Win32 functions). > If the speed increases mentioned are correct, then it certinaly > seems a good proposal. Well maybe I'll have a go at implementing it properly and then trying to get it into Boost. I suppose it can be selectively enabled for those platforms that support it.

1 0

Re: Building boost under mingw: bjam invokes VC7?? (part IV: solu tion)
by Brewer, Christopher 19 Jul '04

19 Jul '04

I have found my way around the problem, though I don't understand why it works. Recap of the problem (for those joining late): attempting to build boost with bjam under msys/mingw resulted in very strange behavior. Within the msys shell, invoking the following line: bjam -sTOOL=mingw install ...resulted in bjam attempting to use the VC7 toolset. I checked the path within msys to make sure I was running the version of bjam.exe that I thought I was. Looking through the directories showed there was indeed only one copy. On a hunch, I moved it from it's present location at 'c:/msys/1.0/bin' to 'c:/msys/1.0/mingw/bin' and tried again. And suddenly bjam began using gcc. 8^0 Apparently where you run it from makes a big difference. I'm sure that an msys guru could take one look at the symptoms and say, "Of course! It's obvious to the meanest intellect!". Not being an msys guru I will just shrug my shoulders and move forward :) All is well now and Boost is building even as I type this message. Many thanks to those who replied! Best regards, Chris Brewer (Company-mandated disclaimer) This message originates from WMS Gaming Inc. It contains information which may be confidential or privileged and is intended only for the individual or entity named above. It is prohibited for anyone else to disclose, copy, distribute or use the contents of this message. All personal messages express views solely of the sender, which are not to be attributed to WMS Gaming Inc., and may not be copied or distributed without this disclaimer. If you received this message in error, please notify us immediately at (847) 785-3000.

2 1

Re: [Boost-users] Installation of boost-sandbox
by Ross Boylan 18 Jul '04

18 Jul '04

On Sun, Jul 18, 2004 at 01:56:20AM -0400, Franz Mueller wrote: > Hi, > I've heared of the ublas-binding to lapack, which sounds very good. > Thus I tried to install it, but I didn't succeed. > I got boost-sandbox using cvs. Then I went to the boost-sandbox directory > and did a bjam. > I'd be interested in hearing the answer to this too, as I've been looking in the sandbox. My impression is that it is not meant to be built as a whole. As the name implies, it's a bunch of separate projects, so I think the intent is that you pull one or more of them out and stick them in a regular boost tree. That's just a guess, and I notice that this seems awkward to do since a single project may appear in several scattered directories.

1 0

array_traits?
by Duane Murphy 18 Jul '04

18 Jul '04

Howdy fellow boosters, Several years ago I started using array_traits. I gathered from some later discussion that this library was never meant to be part of boost. However, I really like the library and find it more useful than the array<> implementation. Point in fact, array_traits and array<> overlap but probably weren't meant to solve the same problems [1]. Is array_traits around any more? I seem to recall something more general purpose. Maybe container_traits? Is this still being actively worked on? I am planning on (finally) upgrading our use of boost to a more recent version. I'm trying to work out what to do about our use of array_traits. Maybe I'll just keep using it. ...Duane [1] The primary difference between array_traits and array<> is the "construction" of the array. array_traits in fact has no construction. It can work with any arbitrary defined C-array of objects. array<> on the other hand requires construction. For my use,the major advantage that array_traits has over array<> is that I can refer to any ordinary C- array; in particular arrays that may not be under my control. Where as array<> requires that I change the type of the object.

3 3

Installation of boost-sandbox
by Franz Mueller 18 Jul '04

18 Jul '04

Hi, I've heared of the ublas-binding to lapack, which sounds very good. Thus I tried to install it, but I didn't succeed. I got boost-sandbox using cvs. Then I went to the boost-sandbox directory and did a bjam. But I got this: libs\numeric\blasbindings\tst\Jamfile: No such file or directory libs\numeric\blasbindings\tst\Jamfile: No such file or directory don't know how to make <@boost-sandbox!libs!numeric!bindings!amos>amos.cpp ...found 227 targets... ...updating 34 targets... ...can't find 1 target... ...can't make 5 targets... MkDir1 bin I am using Visual Studio.NET Prof. 2003. I'd love to try the ublas-bindings, so can anybody help me installing it? Anything else I have to take care of once the bindings work (e.g. link to clapack or so)? -- Franz

1 0

Re: [Boost-users] cvs access problems SOLVED --website update
by Ross Boylan 17 Jul '04

17 Jul '04

On Fri, Jul 16, 2004 at 06:44:28PM -0700, Jeff Garland wrote: > On Fri, 16 Jul 2004 18:24:18 -0700, Ross Boylan wrote > > This happened to me twice: > > $ cvs -d :pserver:anonymous@cvs.boost- > > sandbox.sourceforge.net:/cvsroot/boost-sandbox login Logging in to > > :pserver:anonymous@cvs.boost- > > sandbox.sourceforge.net:2401/cvsroot/boost-sandbox CVS password: # > > I hit enter cvs [login aborted]: connect to cvs.boost- > > sandbox.sourceforge.net(66.35.250.209):2401 failed: Connection timed > > out > > > > Am I doing something wrong (like needing to open up a port in my > > firewall), or is this just a momentary sourceforge lapse? > > > > I've gotten cvs stuff from elsewhere (including other sf projects, I > > think), so I doubt it's my setup. > > Soureforge has removed the project specifc path from url. Note the > cvs.sourceforge.net instead of cvs.boost-sandbox.sourceforge.net : > > $ cvs -d :pserver:anonymous@cvs.sourceforge.net:/cvsroot/boost-sandbox login > > Jeff Thanks, that did it. It might be good to update http://www.boost.org/more/mailing_lists.htm on the bottom, which gives the old location.

3 2