December 2002 - Boost-users

RE: [Boost-Users] Re: boost::date_time : iterator model
by Jeff Garland 28 Dec '02

28 Dec '02

>However I will try to give it a shot. What is the right >place to take latest source code from? Regular Boost CVS >repository on SourceForge? Yes, the boost CVS repository on sourceforge has the latest source. Jeff

1 0

Property maps / Graph library
by rradh <rradh＠yahoo.com> 25 Dec '02

25 Dec '02

Hi, The examples in libs/graph/example have property_map like vertex_name that associates a 'vertex -> user-defined name' (usually a std::string). Without having to use an external std::map, is there any way to create a property_map that provides the reverse association 'user-defined name -> vertex', so that I can retreive the vertex based on the user-defined name and without iterating thru each vertex. Thanks Rajesh

1 0

[BGL] Simplest way to implement bool isReachable(Vertex src, Vert ex target, Graph g) was: [Boost-Users] Re: Extended use of BGL File Depen dency Example
by Tom Matelich 24 Dec '02

24 Dec '02

> From: Tom Matelich [mailto:tmatelich@zetec.com] > > From: Tom Matelich [mailto:tmatelich@zetec.com] > > From: David Abrahams [mailto:dave@boost-consulting.com] > > > Tom Matelich <tmatelich(a)zetec.com> writes: > > > > From: David Abrahams [mailto:dave@boost-consulting.com] > > > >> Tom Matelich <tmatelich(a)zetec.com> writes: > > > >> > > > >> > trying to make my graphviz representation look nicer and I > > > >> was wondering if > > > >> > there is an algorithm already for simplifying: > > > > > > > > -----B\ > > > > A--< \ > > > > --------C > > > > D----------/ > > > > > > > > to: > > > > > > > > A---B---C > > > > D------/ > > > > > > > > This eliminated the A-C edge. <snip discussion of minimum spanning tree. It won't work because I require all reachability to be maintained. > so here's my psuedo-algorithm: for each vertex V for each out-edge E for each out-edge E_prime (double loop) if(isReachable(target(E_prime), target(E)) remove_edge(E) In other words if one of V's children can reach the target vertex of a given edge, then that edge can be removed. If someone could let me know how to implement bool isReachable(Vertex src, Vertex target, Graph g); I would really appreciate it. Or if you have a better algorithm, I'd love to hear it. Thanks, Tom ----------------------------------------------------------------------- DISCLAIMER: Information contained in this message and/or attachment(s) may contain confidential information of Zetec, Inc. If you have received this transmission in error, please notify the sender by return email. -----------------------------------------------------------------------

1 0

RE: [Boost-Users] Re: Extended use of BGL File Dependency Example
by Tom Matelich 23 Dec '02

23 Dec '02

> From: Tom Matelich [mailto:tmatelich@zetec.com] > Sent: Sunday, December 01, 2002 9:11 PM > > From: David Abrahams [mailto:dave@boost-consulting.com] > > Tom Matelich <tmatelich(a)zetec.com> writes: > > > > > From: David Abrahams [mailto:dave@boost-consulting.com] > > > Sent: Sunday, December 01, 2002 5:58 AM > > >> Tom Matelich <tmatelich(a)zetec.com> writes: > > >> > > >> > trying to make my graphviz representation look nicer and I > > >> was wondering if > > >> > there is an algorithm already for simplifying: > > > > > > -----B\ > > > A--< \ > > > --------C > > > D----------/ > > > > > > to: > > > > > > A---B---C > > > D------/ > > > > > > This eliminated the A-C edge. > > > > > > Thanks for considering my problem. > > > > Maybe you want a minimum spanning tree, where the weight of > > each edge in the tree is determined by the distance between its > > vertices in some topological sort? > > > > That seems to solve the cases shown above pretty nicely. > > I had looked at the minimum spanning tree docs a bit, I > thought it looked pretty good. The docs say undirected graph though. I'm > currently using a directed graph. I suppose I could create a separate graph. > Oh, I see, create an undirected graph with weights as you said, and use > the minimum spanning algorithm to choose the useful edges. I like it, > hope it works. > Well, I finally got back to this and it almost worked. Things look a lot nicer, the only problem is some edges were lost that I would want to keep. I'll give a list of edges rather than try to draw ascii graphs. So I start out with edges: [1,0], [2,0], [3,0], [4,0], [3,1], [4,2], [4,3] Ideally I want this reduced to: [1,0], [2,0], [3,1], [4,2], [4,3] What I'm getting is: [1,0], [3,1], [4,2], [4,3] Of course my graph is bigger and I have a lot more dead ends than just at 2. 0 is a place holder I have for a "Valid System", so every thing should hit that eventually. I implemented this in the following manner: [create typedefs for undirected graph to be minimized] [get the order from the topological sort in a vector] [create the undirected graph with all edges from dependency graph, assigning weight to be the distance between source and target in the topological sort] [create a vector of edges called spanning_tree] [call kruskal_minimum_spanning_tree(undirected_graph, std::back_inserter(spanning_tree)); ] [then I recreated the directed graph out of the edges in spanning_tree, so I could have my arrows in graphviz] I'm going to try using prim_minimum_spanning_tree, I assume I should expect the same results, but I'll see. Thanks for any advice, Tom ----------------------------------------------------------------------- DISCLAIMER: Information contained in this message and/or attachment(s) may contain confidential information of Zetec, Inc. If you have received this transmission in error, please notify the sender by return email. -----------------------------------------------------------------------

1 0

Error in testing class tuple
by agiatojon <JonAgiato＠nyc.rr.com> 23 Dec '02

23 Dec '02

Happy Holidays to everyone! I am getting aquainted with the tuple class but am encountering the following error: c:\BGL\boost_1_29_0\boost\tuple\tuple_comparison.hpp(119): error C2027: use of undefined type 'boost::STATIC_ASSERTION_FAILURE<x>' with [ x=false ] with the following program: /** * tupleTest.cpp * Test of Boost library class tuple * A tuple is a fixed size collection of elements * Stored objects may be of different types * Current support of tuples with 0-10 objects * Objects may be any valid C++ type except void and function types * Jon Agiato * 12-23-02 */ #include <iostream> using std::cout; using std::cin; using std::ostream; using std::cerr; #include <cstdlib> #include <boost/tuple/tuple.hpp> using boost::tuple; // both tuple_io.hpp and tuple_comparison.hpp include tuple.hpp // tuple_io must be included for tuple input and output operators #include <boost/tuple/tuple_io.hpp> // tuple_comparison.hpp must be included for comparison operators #include <boost/tuple/tuple_comparison.hpp> #include <string> using std::string; #include <exception> using std::exception; class A { friend ostream &operator<<(ostream &output, const A &a) { output << a.S; return output; } public: A(string s = "Testing"):S(s) {;} string S; }; class B { friend ostream &operator<<(ostream &output, const B &b) { output << b.D; return output; } public: B(double d = 3.501):D(d) {;} double D; }; class C { friend ostream &operator<<(ostream &output, const C &c) { output << c.state; return output; } public: C(bool st = true):state(st) {;} double state; }; int main() { try { tuple<A, B, C> testTuple; cout << testTuple << '\n'; tuple<B, A> test2Tuple; if(testTuple == test2Tuple) cout << testTuple << " is equal to " << test2Tuple << '\n'; else cout << testTuple << " is not equal to " << test2Tuple << '\n'; cout << "Press enter to exit..."; cin.get(); return EXIT_SUCCESS; } catch(const exception &E) { cerr << "Standard library exception: " << E.what() << '\n'; return EXIT_FAILURE; } catch(...) { cerr << "Unknown exception thrown.\n"; return EXIT_FAILURE; } } I am using Visual Studio.NET Enterprise Architect Thanks! Jon Agiato JonAgiato(a)nyc.rr.com

1 0

Newbie: customizing dfs
by Gil Shallom 23 Dec '02

23 Dec '02

Hi, Sorry if I missed anything in any FAQ. I would like to customize dfs so it would go over an edge only if satisfies certain properties. The library allows to take actions using visitors, but I haven't found a nice way to modify its decision so it would scan only part of the tree according to certain properties. I can change depth_first_visit_impl in depth_first_search.hpp or maybe define a new algorithm in a new namespace, but all I actualy want is something like my visitor returning false/true thus telling the algorithm whether to continue scanning an edge or not. Does anyone think of a nice way to do this? Or have I missed a feature already implemented in BGL? Any help would be appreciated, Thanks, Gil Shallom

1 0

boost::python : project outside boost tree
by Juerg Tschirren 22 Dec '02

22 Dec '02

I have trouble getting a boost::python project to compile outside the boost directory tree. I can compile the "hello, world" example in {boost}/libs/python/example/tutorial witout problem. But now I would like to start my own project outside the boost tree. To start with I copied the above "tutorial" directory into my home-directory and modified the Jamfile to: #--- begin Jamfile # Include definitions needed for Python modules SEARCH on python.jam = $(BOOST_BUILD_PATH) ; include python.jam ; extension hello : hello.cpp <dll>$(BOOST_ROOT)/libs/python/build/boost_python : <sysinclude>$(BOOST_ROOT) ; #---- end Jamfile I set export PYTHON_ROOT=/usr export PYTHON_VERSION=2.2 export BOOST_BUILD_PATH=/usr/lib/boost_1_29_0 export BOOST_ROOT=/usr/lib/boost_1_29_0 (boost is in /usr/lib/boost_1_29_0 and Python 2.2.1 is in /usr/lib/python2.2 and /usr/include/python2.2) When I now call bjam I get the messages: ../../../Jamrules: No such file or directory /usr/lib/boost/tools/build/boost-base.jam:1455: in find-compatible-subvariant *** argument error * rule is-link-compatible ( feature : value1 : value2 ) * called with: ( <target-type> : PYD : ) * missing argument value2 /usr/lib/boost/tools/build/boost-base.jam:1416:see definition of rule 'is-link-compatible' being called /usr/lib/boost/tools/build/boost-base.jam:1502: in link-libraries /usr/lib/boost/tools/build/boost-base.jam:1712: in subvariant-target /usr/lib/boost/tools/build/boost-base.jam:1797: in main-target /usr/lib/boost/tools/build/boost-base.jam:1330: in declare-local-target /usr/lib/boost/tools/build/python.jam:302: in extension Jamfile:29: in load-jamfiles /usr/lib/boost/tools/build/bootstrap.jam:15: in boost-build /usr/lib/boost/boost-build.jam:2: in module scope I suspect something is wrong with my Jamfile, but what? Any help would be greatly appreciated. Juerg

2 3

[BGL] Possible bug in minimum_degree_ordering
by Jeppe N. Madsen 22 Dec '02

22 Dec '02

Hi, I'm fairly new to BGL so this may not be a bug after all......but I get a fair number of warnings about possible loss of data due to conversions (mainly between signed/unsigned types) with MSVC7 trying to compile the example below. Some may be ignored but at least line 188 of minimum_degree_ordering.hpp seems wrong: data[id[i]] = - (id[j] + offset); id[j] is unsigned, making (id[j] + offset) unsigned thus making the unary minus rather spurious. Also, compile time checks for correctly using signed types for the permutations would be nice :-) rgds Jeppe #include <iostream> #include <boost/graph/adjacency_list.hpp> #include "boost/graph/minimum_degree_ordering.hpp" #include <iostream> #include <boost/graph/adjacency_list.hpp> #include "boost/graph/minimum_degree_ordering.hpp" using namespace std; using namespace boost; int main(int, char **) { typedef adjacency_list <vecS, vecS, directedS> Graph; Graph G; add_edge(0, 1, G); add_edge(1, 0, G); add_edge(0, 2, G); add_edge(2, 0, G); add_edge(0, 5, G); add_edge(5, 0, G); add_edge(1, 2, G); add_edge(2, 1, G); add_edge(1, 3, G); add_edge(3, 1, G); add_edge(1, 6, G); add_edge(6, 1, G); add_edge(2, 4, G); add_edge(4, 2, G); add_edge(3, 4, G); add_edge(4, 3, G); add_edge(4, 5, G); add_edge(5, 4, G); std::vector<size_t> degree(num_vertices(G), 0); std::vector<int> inverse_perm(num_vertices(G), 0), perm(num_vertices(G), 0); std::vector<size_t> supernode_sizes(num_vertices(G), 1); property_map<Graph, vertex_index_t>::type id = get(vertex_index, G); minimum_degree_ordering (G, make_iterator_property_map(&degree[0], id, degree[0]), &inverse_perm[0], &perm[0], make_iterator_property_map(&supernode_sizes[0], id, supernode_sizes[0]), 0, id); for(size_t i = 0; i < num_vertices(G); ++i) std::cout << inverse_perm[i] << "\n"; return 0; }

1 0

boost::python abstract base class
by John Hunter 20 Dec '02

20 Dec '02

I am trying to get raw pointers and abstract base classes working properly with boost::python v2. I am working on an example inspired by the tutorial http://www.boost.org/libs/python/doc/tutorial/doc/inheritance.html. In my case, the base class has a pure virtual function, but I don't need to override it in python and call it from c++, so I don't think I need to do a BaseWrap. The problem is that if I try and expose Base to python, I get the compiler (gcc 3.2) error: g++ -isystem/usr/local/include -I/usr/local/include/python2.2 -fPIC -ftemplate-depth-100 -c rawpointer.cpp /usr/local/include/boost/python/object/value_holder.hpp: In instantiation of `boost::python::objects::value_holder<Base>': /usr/local/include/boost/type_traits/alignment_of.hpp:52: instantiated from `boost::detail::alignment_of_impl<boost::python::objects::value_holder<Base> >' /usr/local/include/boost/python/object/instance.hpp:29: instantiated from `boost::alignment_of<boost::python::objects::value_holder<Base> >' ...snip... rawpointer.cpp:47: instantiated from here /usr/local/include/boost/python/object/value_holder.hpp:36: cannot declare field `boost::python::objects::value_holder<Base>::m_held' to be of type ` Base' /usr/local/include/boost/python/object/value_holder.hpp:36: because the following virtual functions are abstract: rawpointer.cpp:15: virtual const std::string& Base::get_name() rawpointer.cpp:14: virtual void Base::sayit() const make: *** [rawpointer.o] Error 1 and if I only expose the derived classes, the code compiles but then I get a runtime error when I try and import the class in python >>> import rawpointer Traceback (most recent call last): File "<stdin>", line 1, in ? RuntimeError: extension class wrapper for base class 4Base has not been created yet >>> Below is my example code. Suggestions welcomed! John Hunter // work with raw pointers #include <iostream> #include <string> #include <boost/python.hpp> #include <boost/python/manage_new_object.hpp> using namespace boost::python; class Base { public: virtual ~Base() {}; virtual void sayit() const=0; virtual const std::string& get_name()=0; private: }; class Foo : public Base { public: Foo() : _name("Foo") {} void sayit() const { std::cout << "Foo" << std::endl; } const std::string& get_name() {return _name;} private: const std::string _name; }; class Bar : public Base { public: Bar() : _name("Bar") {} void sayit() const { std::cout << "Bar" << std::endl; } const std::string& get_name() {return _name;} private: const std::string _name; }; Base* barfactory(){ return new Bar(); } Base* foofactory(){ return new Foo(); } std::string simonsays(Base* p) {return p->get_name();} BOOST_PYTHON_MODULE_INIT(rawpointer) { class_<Base>("Base", no_init) .def("sayit", &Base::sayit); class_<Foo, bases<Base> >("Foo") .def("sayit", &Foo::sayit); class_<Bar, bases<Base> >("Bar") .def("sayit", &Bar::sayit); def("foofactory", foofactory, return_value_policy<manage_new_object>()); def("barfactory", barfactory, return_value_policy<manage_new_object>()); def("simonsays", simonsays); }

2 2

Why file_system build on msvc 6.5 is only dynamic?
by Daniel Yerushalmi 20 Dec '02

20 Dec '02

Hi All I got filesyetm from the CVS and build it using bjam (no args with VCVARS32.BAT of msvc 6.5) Only the folder: runtime-link-dynamic was created under bin/lisbfs.lib/msvc. Other libraies (i.e. thread) build by default also runtime-link-static. What should I do to do the build with static run time library? Thanks Daniel

3 5