October 2002 - Boost-users - lists.boost.org

boost::shared_ptr and stl algorithms
by Bill Eidson 15 Apr '05

15 Apr '05

std::vector< boost::shared_ptr<T> > TVec; I want to write a function object to run std::for_each. to do somework against members of T and accumulate results in functor. My attempts have failed in VC7 with memory errors in my operator() method. struct fo { bool operator()( boost::shared_ptr<T>& t ) { return true; } }; Thanks in advance bille

5 9

BGL and reversing / moving edges
by joshrose 17 Dec '03

17 Dec '03

G'day, I was hoping someone might be able to tell me if they've run into any situations using BGL where they needed to move (or reverse (i.e. flip the source and target)) an edge within a graph implemented as an adjacency_list. I'm working on an implementation of an algorithm for max. weight bipartite matching ("the assignment problem"), and one of the assumptions in the algorithm is that non-matched edges are directed from one subgraph, A, to the other subgraph, B, and matched edges are directed the other way. Over the course of growing the matching, a simple way to add and remove edges from the candidate set is just to reverse the edge, thus the need for some way to reverse the edge within BGL. I don't think I can just remove and then re-add the edge in the other direction, primarily because doing so will (I believe) invalidate any edge_descriptors referring to that edge. In particular, there are property maps keeping track of the weight of edges, the predecessors of certain nodes, etc. that use edge descriptors as keys. I've spent a fair bit of time reading the documentation and looking through the implementation of adjacency_lists (incl. edge_list_impl), and I'm not really sure of a way to either move an edge directly within a graph, or even to create a new edge from another s.t. they have the same edge_descriptor (in general a bad thing, probably). Anyone have any thoughts on this? Thanks, -Josh

4 10

bjam and windows command line problem.
by Fernando Cacciola 07 May '03

07 May '03

Hi, I'm having a hard time trying to use bjam to build boost 1.29.0. I have Win98 in Spanish, and I use Borland C++ Builder 5.0 (its bcc32.exe corresponds to is BCC5.5.1) I have boost 1.29.0 installed at: "C:\Libraries\boost\Repository\boost" I've downloaded bjam for Microsoft Windows: >bjam -v Boost.Jam Version 3.1.0. OS=NT. Copyright 1993-2002 Christopher Seiwald and Perforce Software, Inc. Copyright 2001 David Turner. Copyright 2001-2002 David Abrahams. I've temporarily modified the root 'jamfile' so that it builds only the Test library. and I've done: subst d: "c:\archivos de programa\borland\cbuilder5" set BCCROOT=d: (you'll see why in a moment) If I try: c:\>cd C:\Libraries\boost\Repository\boost" C:\Libraries\boost\Repository\boost> bjam -sTOOLS=borland I receive messages of the form: borland-C++-action libs\test\build\bin\libpem.lib\borland\debug\runtime-link-static\threading-s ingle\execution_monitor.obj "d:\bin\bcc32" -j5 -q -c -P -w -Ve -Vx -a8 -v -Od -vi- -tWC -WM- -w-8001 -I"libs\test\build" -I"C:\Libraries\boost\Repository\boost" -I"d:\includ e" -o"libs\test\build\bin\libpem.lib\borland\debug\runtime-link-static\thre ading-single\execution_moni Error E2266: No file names given tor.obj" "libs\test\build\../src\execution_monitor.cpp" C:\WINDOWS\TEMP\JAMTMP00.BAT [2] Unknown command "TOR.OBJ" "LIBS\TEST\BUILD\.." "d:\bin\bcc32" -j5 -q -c -P -w -Ve -Vx -a8 -v -Od -vi- -tWC -WM- -w-8001 -I"libs\test\build" -I"C:\Libraries\boost\Repository\boost" -I"d:\includ e" -o"libs\test\build\bin\libpem.lib\borland\debug\runtime-link-static\thre ading-single\execution_monitor.obj" "libs\test\build\../src\execution_monitor.cpp" ...failed borland-C++-action libs\test\build\bin\libpem.lib\borland\debug\runtime-link-static\threading-s ingle\execution_monitor.obj... If I'm interpreting this output correctly, the problem is that the command line feed to bcc32 exceeds 256 characters. What can be done about this? Is there a way for 'borland-tools.jam' to put the bcc32 parameters in a tmp response file and use: bcc32.exe @tmprsp.rsp? Any help will be appreciated. TIA, -- Fernando Cacciola

4 22

VC7 says 'fatal C1067' compiling multi_array?
by Clift Norris 01 Mar '03

01 Mar '03

Greetings: I'm using VC7 and boost_1_29_0 to compile this code that is using multi_array.hpp. typedef boost::multi_array<double, 4> theta4_t; typedef theta4_t::index index; theta4_t Theta_Trun( boost::extents[10000][2][10][256] ); I'm receiving this error: "fatal error C1067: compiler limit: debug information module size exceeded" Of course, the code compiles fine using the "Release" configuration, but I need the "Debug" one also. Is there a workaround for this, other than the 'make your variables shorter' suggested in the VC7 documentation? Thanks, -Clift

2 1

bug report: boost::tokenizer with VC6
by Navi Singh 09 Feb '03

09 Feb '03

The boost release 1.28.0 of the tokenizer library has the following problem: There are 5 examples, namely, simple_example_1.cpp thru simple_example_5.cpp If I change the main() in each of the 5 to to main1() thru main5() and return 0; the vc60 compiler generates an internal compiler error when parsing the file boost\iterator_adapters.hpp on line 869, where it is parsing the following lines: reference operator*() const { return policies().dereference(*this); }

3 2

Q: Graphs. Sorting edges. Possible? How to?
by stas_fomin 07 Nov '02

07 Nov '02

Hello. I a newbie in BGL and I rewrite my LEDA code now... I have to sort (order) edges according to some functor. For example (simplified): typedef adjacency_list<vecS, vecS, bidirectionalS> MyGraph; typedef graph_traits < MyGraph >::edge_descriptor MyGraphEdge; typedef graph_traits < MyGraph >::edge_iterator MyEdgeIter; class MyCompareEdges{ public: bool operator()(MyGraphEdge e1, MyGraphEdge e2) const { return true; } }; .... { MyGraph Gr(10); MyEdgeIter ei_beg, ei_end; MyCompareEdges comp; tie(ei_beg, ei_end) = edges(Gr); std::sort(ei_beg, ei_end, comp); //Error C2676 } And I have error C2676 (below the message...). Questions: 1. Is it possible to sort edges in the graph? 2. How to... Thanks all in advance... Stas Fomin. P.S. vecS, setS, listS for edge container... - C2676. ------------------------------ d:\projects\lib\cpp_libs\libboost\boost\iterator_adaptors.hpp(264) : error C2676: binary '+=' : 'class std::list<struct boost::list_edge<unsigned int,struct boost::no_property>,class std::allocator<struct boost::list_edge<unsigned int,struct boost:: no_property> > >::iterator' does not define this operator or a conversion to a type acceptable to the predefined operator d:\projects\lib\cpp_libs\libboost\boost\iterator_adaptors.hpp (911) : see reference to function template instantiation 'void __thiscall boost::default_iterator_policies::advance(struct boost::iterator_adaptor<class std::list<struct boost::lis t_edge<unsigned int,struct boost::no_property>,class std::allocator<struct boost::list_edge<unsigned int,struct boost::no_property> > >::iterator,struct boost::detail::undirected_edge_iter_policies,class boost::detail::edge_desc_impl<struct boost::b idirectional_tag,unsigned int>,class boost::detail::edge_desc_impl<struct boost::bidirectional_tag,unsigned int>,class boost::detail::edge_desc_impl<struct boost::bidirectional_tag,unsigned int> *,struct boost::multi_pass_input_iterator_tag,int> &,i nt)' being compiled Error executing cl.exe.

4 7

Q: Problems building boost.python under cygwin
by "Schnörr, Claudius Dr." 06 Nov '02

06 Nov '02

Hello, I run into problems building boost under cygwin. It seems to be merely a matter of a wrong configuration or wrong scripts. After running 'bjam -sTOOLS=gcc -n > build.com' in boost/libs/python/build in order to get hands on the statements being executed by bjam, there had to be made a lot of textual replacements to get the statements in build.com run: Just to mention some: - replace '-isystem ' by '-I' - remove '-I""' - change shared library-names from *.so to *.dll - insert '-L <path_to_libpython2.2.dll> -lpython2.2' where shared libs are created In boost/libs/python/test, going the same way, there have in addition this changes to be made: - LD_LIBRARY_PATH=... --> PATH=$PATH:..., because cygwin uses PATH to look for .dlls My question: I think there have to be made more configurations than I did to create the boost.python-stuff under cygwin. Where have they to be done, and what? (I already configured environment variables like GCC_ROOT_DIRECTORY, GXX, PYTHON_ROOT, ...etc...) Any suggestions are welcome! Please send a copy of your answer to schnoerr(a)mailzone.com Thank you in advance, Claudius --------------------------------------------------------- This Mail has been checked for Viruses Attention: Encrypted mails can NOT be checked! ** Diese Mail wurde auf Viren geprueft Hinweis: Verschluesselte mails koennen NICHT auf Viren geprueft werden! ---------------------------------------------------------

2 1

Wrapping a std::list with boost.python
by jefflholle 03 Nov '02

03 Nov '02

I've just started experimenting with version 2 of boost.python. In iterator.cpp, a std::list is wrapped using following code: typedef std::list<int> list_int; void push_back(list_int& x, int y) { x.push_back(y); } int back(list_int& x) { return x.back(); } . . . class_<list_int>("list_int") .def("push_back", push_back) .def("back", back) .def("__iter__", iterator<list_int>()) ; Is it possible to use the container methods instead of user defined functions? I've tried the code: class_<list_int>("list_int") .def("push_back", &list_int::push_back) .def("back", &list_int::back) .def("__iter__", iterator<list_int>()) ; This doesn't work. Am what I'm tried to do possible? If so, how?

2 1

[BGL] How to prune a BFS or exit a BFS
by kweeheong 01 Nov '02

01 Nov '02

I have a custom BFS visitor used to traverse a graph. Is there a built-in way to - prune off a vertex's descendants from the search - exit a BFS These ( pruning and terminating ) are separate situations. I believe the following will work; however, I do not know how to access the relevant data structures. - To prune off a vertex's descendents from the search In the custom BFS visitor's discover_vertex event, set the color of all its adjacent descendants to "black". - To exit a BFS In the custom BFS visitor's discover_vertex event, once the required vertex is found, empty the queue of vertices. How can I access a vertex's color and a BFS's queue? Or is there a better way?

2 2

How to pass result of boost::bind to arg of template function ?
by hugo duncan 31 Oct '02

31 Oct '02

Hi, I want to pass the result of a boost::bind as an argument of a templated function of a class, and am struggling to find a signature that will let me do this. Is this possible? if so what signature should the function have ? Any help much apreciated Hugo #include "boost/function/function1.hpp" #include "boost/bind.hpp" class another_class {}; class my_class { public: typedef another_class ObjectType; //! this works fine passing a boost::function object template <class ReturnType> void add(const boost::function1<ReturnType,ObjectType>& arg) { boost::function1<ReturnType,ObjectType> fn_obj=arg; } //! this should work when passed the result of a boost::bind template <class ReturnType> void add(?what goes here? & arg) { function1<ReturnType,ObjectType> fn_obj(arg); } }; int fn(another_class,int) { return 1; } int main(int, char**) { my_class c; c.add(boost::bind(&fn,_1,1)); }

2 1

[bgl] breadth_first_search for shortest paths (fwd)
by Joel Young 31 Oct '02

31 Oct '02

Dear Jeremy, et al, I am trying to implement a shortest paths algorithm on a directed graph. As I have unit edge weights, I would rather use breadth_first_search rather than dijkstra. // Is this a good candidate for addition? template <class Map, class Tag> struct vertex_marker : public boost::base_visitor<vertex_marker<Map,Tag> > { typedef Tag event_filter; vertex_marker(Map p_map,int p_val) : m_map(p_map),m_val(p_val) { } template <class Vertex, class Graph> void operator()(Vertex v,const Graph&) { boost::put(m_map,v,m_val); } Map m_map; int m_val; }; template <class Map, class Tag> vertex_marker<Map,Tag> mark_vertices(Map p_map, Tag, int p_v) { return vertex_marker<Map,Tag>(p_map,p_v); } boost::breadth_first_search( mygraph, rootnode, boost::visitor(boost::make_bfs_visitor( std::make_pair( boost::record_distances( boost::get(boost::vertex_distance,mygraph), boost::on_tree_edge()) , std::make_pair( mark_vertices( boost::get(boost::vertex_distance,mygraph), boost::on_initialize_vertex(), std::numeric_limits<int>::max()) , mark_vertices( boost::get(boost::vertex_distance,mygraph), boost::on_start_vertex(), 0) ) ) )) ); The problem is that on_start_vertex does not seem to be implemented for breadth_first_search, but it is for depth_first. For dijkstra, the start node seems to be automaticly zeroed. I guess I could mod Am I missing something really obvious here or is there an inconsistency across the algorithms? Shouldn't on_start_vertex be used for all of the graph algorithms? Also, I can't seem to find documentation on neighbor_bfs.hpp Thanks, Joel

2 1

Debug regex++ in VC6
by r_gharavi 31 Oct '02

31 Oct '02

When I build a stand-alone debug xll with regex++ (static linking), Excel refuses to accept it as an addin. It works fine in release mode. Due to reasons beyond my control :-), I am using VC6 STL. Any pointers would be appreciated. Thanks.

1 0

d_heap bug
by Tsvetomir Valtchev 31 Oct '02

31 Oct '02

[I tried to contact the library's author, dietmar.kuehl@claas- solutions.de, but the message bounced back.] Dear Dietmar, I was looking for a better heap class than the one in STL and I was delighted to come across your d_heap class included in the Boost library. Unfortunately, I had a number of problems with it, and I figured you might be interested in some feedback. First, the class didn't compile with VC.NET (and probably won't with VC 6 either). This might be due to non-conformance on the part of Microsoft's compiler and some known limitations with template support, but for some of the issues I think they were due to the non- conformance of the code itself. Second, to my dismay, the code turned out to be seriously flawed. The following simple fragment, for example, causes a crash: THeap::pointer ptrs[10]; ptrs[0] = heap.push(12); ptrs[1] = heap.push(16); heap.remove(ptrs[0]); heap.remove(ptrs[1]); After some debugging and deciphering of the code, I found that the remove() function didn't handle a special case where the node to be 'sifted up' coincided with the last "hole" left in the tree. I think pop() had a similar problem. (In my fixed version, I have expressed pop() in terms of remove(), which resolves the problem and also avoids some code duplication.) I hope some day you get around to publishing an improved version. Another suggestion I have is some way to make sure that the pointer's passed back by client code are valid. Extensive use of assertions throughout the code might also prevent various programming errors. Regards, // Chango V.

1 0

Regex++ and UTF-8
by Gitit 31 Oct '02

31 Oct '02

Hi, We want to use regex++ (version 3.31) with UTF-8 strings. I tried to match a UTF-8 character of 2 bytes to the regex "." and the match failed. It seems regex++ handles these 2 bytes as two separate characters. 1) Is there a "native" way in the regex++ library for using UTF-8 strings? Can we use UTF-8 strings to compare against a compiled regex (the regex is in ASCII only)? Can the regex itself hold UTF-8 characters? 2) Is converting to wchar_t our only option? As far as I understand, wchar_t does not cover the entire range of characters covered by UTF-8, so it may not be enough. Any other ideas? thanks, Gitit.

1 0

Re: [Boost-Users] Re: regex: flag for "global" search?
by dick.bridges＠tais.com 30 Oct '02

30 Oct '02

The boost::regex_search example bumps the start to march through the sub-expressions and the boost::regex_grep examples demonstrate using the 'callback' for each term. I was kinda hoping the library provided a "convenience version" that would take something like "(\\w+)(?:\\.(\\w+))*" and populate the match_results as if the terms had been expanded <e.g.> "(\\w+)(?:\\.(\\w+))?(?:\\. (\\w+))?(?:\\.(\\w+))?"</e.g.> Perhaps to some arbitrary limit? Oh well, given the riches provided, a little while loop now and then is surely a small price to pay. ;) Thanks for the response. I'm continually amazed at the ethos fostered by Boost group members. "Edward Diener" <eddielee@tropic To: boost-users(a)yahoogroups.com soft.com> cc: Sent by: news Subject: [Boost-Users] Re: regex: flag for "global" search? <news(a)main.gmane .org> 10/29/2002 04:54 PM Please respond to Boost-Users Use boost::regex_grep. The boost::regex_search only finds the first match. <dick.bridges(a)tais.com> wrote in message news:OF6181960C.1A0F5C63-ON88256C61.006996C6@tais.net... > I'm trying to capture the sub-expressions in a fqdn (e.g., > one.two.three.com.) One piece of Perl-speak for the expression is: "/ > (\w+)\./g". I've been using boost_search and trying variations on "(?: > (\\w+)\\.)+" but I can't seem to get all the sub-expressions into the > match_results. > > There's got to be a flag I've been overlooking. Any suggestion will be > appreciated. Info: <http://www.boost.org> Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

1 0

regex: flag for "global" search?
by dick.bridges＠tais.com 30 Oct '02

30 Oct '02

I'm trying to capture the sub-expressions in a fqdn (e.g., one.two.three.com.) One piece of Perl-speak for the expression is: "/ (\w+)\./g". I've been using boost_search and trying variations on "(?: (\\w+)\\.)+" but I can't seem to get all the sub-expressions into the match_results. There's got to be a flag I've been overlooking. Any suggestion will be appreciated. TIA

3 2

Gcc-3.2, statistic_tests.cpp and <limits>
by pcarlini.rm 29 Oct '02

29 Oct '02

Hi, the PR libstdc++/6718 in the Gcc bug database is about a run time problem with this Boost testcase: paolo:~/Programming/c++/boost_1_29_0/libs/random> a.out Confidence level: 0.99; 1-alpha = 0.01; chi_square(19, 36.191) = 0.99 Running tests on minstd_rand KS: 4750* [0] 9500* [0] equidistribution: a.out: statistic_tests.hpp:571: void generic_counter<RandomA ccessContainer>::operator()(int) [with RandomAccessContainer = std::vector<int, std::allocator<int> >] : Assertion `static_cast<unsigned int>(i) < container.size()' failed. Abort According to Nathan Sidwell and Gaby Dos Reis analysis the problem is due to well known issues with numeric_limits<double>::infinity of 3.2's run time library for which unexpectedly Boost provides no work arounds. On the contrary, 2.95.x worked just fine. Please let me know if you need any additional information, ciao, Paolo.

1 1

AW: [Boost-Users] boost::bind problem
by Jasper, Alexander 29 Oct '02

29 Oct '02

-----Ursprüngliche Nachricht----- Von: Mark Snelling [mailto:mark@bakedbeans.com] Gesendet: Dienstag, 29. Oktober 2002 17:44 An: Boost-Users(a)yahoogroups.com Betreff: [Boost-Users] boost::bind problem [...] typedef priority_queue<shared_ptr<Functor>, vector<shared_ptr<Functor> >, bind(less<int(), bind(&Functor::priority, _1), bind(&Functor::priority, _2))> MyQueue; As far as I know you can't call a function in a typedef. bind(..) is a function call returning a predicate. But does s.o. know how to use bind in this manner. I failed to, since the type of what bind returns is far to complex. And it is required if you want to to use bind in conjunction with maps or the priority_queue's. You have to wrap the call in your own predicate. alex

1 0

boost::bind problem
by Mark Snelling 29 Oct '02

29 Oct '02

Hello, I was wondering if anyone out there in Boost land could give me a hand here? Here's the problem: I have a class A: class A { public: A(int i) : _priority(i) {} int priority() {return _priority;} private: int _priority; }; What I want to do is create a std::priority_queue based on boost::shared_ptr<A> and use boost::bind/compose to create the sorting function required by std::priority_queue. However, when I try and typedef the queue (to save a lot of typing) I get compiler errors on VS.NET. using namespace std; using namespace boost; typedef priority_queue<shared_ptr<Functor>, vector<shared_ptr<Functor> >, bind(less<int(), bind(&Functor::priority, _1), bind(&Functor::priority, _2))> MyQueue; Am I going about this the right way? Any help appreciated. TIA Mark.

1 0

Graph library: parallel edges
by Juerg Tschirren 29 Oct '02

29 Oct '02

With the Boost Graph Library, I try do create an undirected graph that does not allow parallel edges, so far without success. Does anybody have a small example that shows how to do this? Juerg

2 1

Graph library: vertex type
by Juerg Tschirren 29 Oct '02

29 Oct '02

In the Boost Graph Library, is it possible to have a custom-defined type for vertices, e.g., I would like to have boost::graph_traits<Graph>::vertex_descriptor map to std::pair<int,int> Can this be done? If yes, how? Juerg

2 1

Re: [Boost-Users] MSVCNT not used anymore?
by dick.bridges＠tais.com 29 Oct '02

29 Oct '02

Thanks! That worked. The MSVCNT example on http://www.boost.org/tools/build/build_system.htm must be left over from a previous version. Appreciate the quick response. sashan <sashang(a)ihug.co To: Boost-Users(a)yahoogroups.com .nz> cc: Subject: Re: [Boost-Users] MSVCNT not used anymore? 10/29/2002 04:20 AM Please respond to Boost-Users dick.bridges(a)tais.com wrote: >I'm having a senior moment. > >MSVCNT is set to C:\Program Files\Microsoft Visual Studio\VC98 and bjam is >invoked with "-sTOOLS=msvc" but the command line that gets invoked is C: >\Program Files\Microsoft Visual C++\VC98\bin\cl and, of course I get: "The >system cannot find the file specified." > >Where's the instructions for compiling with VC6, SP4? > > This is my command line (saved in a batch file): bjam -sTOOLS=msvc -sBUILD=release -sMSVC_ROOT=D:\msvs\VC98 -sVISUALC=D:\msvs\VC98 I don't think MSVCNT is used see http://www.boost.org/tools/build/msvc-tools.html -- P: Egad! You astound me, Brain! B: That's a simple task, Pinky. Info: <http://www.boost.org> Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

1 0

MSVCNT not used anymore?
by dick.bridges＠tais.com 29 Oct '02

29 Oct '02

I'm having a senior moment. MSVCNT is set to C:\Program Files\Microsoft Visual Studio\VC98 and bjam is invoked with "-sTOOLS=msvc" but the command line that gets invoked is C: \Program Files\Microsoft Visual C++\VC98\bin\cl and, of course I get: "The system cannot find the file specified." Where's the instructions for compiling with VC6, SP4?

2 1

are examples on maual page ready to run?
by J-S Lin 28 Oct '02

28 Oct '02

The error message is as follows, --------------------------------- g:\C++ notes\test_boost\test_matrix.cpp(14) : error C2679: binary '<<' : no operator found which takes a right-hand operand of type 'boost::numeric::ublas::matrix' (or there is no acceptable conversion) with [ T=double, F=boost::numeric::ublas::row_major, A=boost::numeric::ublas::unbounded_array ] ------------------------------ Much appreciated. J-S Lin

2 2

lexical_cast fails to convert std::string to boost::dynamic_bitset<>
by Klaus Nowikow 28 Oct '02

28 Oct '02

Hi! I am trying to convert a string to a boost::dynamic_bitset using boost::lexical_cast, as in this example: template <class Value> bool GetValue(const std::string& str, Value& val) { try { val = boost::lexical_cast<Value>(str); } catch(boost::bad_lexical_cast&) { return false; } return true; } int main() { std::string str = "01"; boost::dynamic_bitset<> bits; if(!GetValue(str, bits)) { [...] } } The lexical_cast always throws an exception. The problem seems to be in lexical_cast.hpp, where a temporary object is created: Target result; if(!(interpreter << arg) || !(interpreter >> result) || !(interpreter >> std::ws).eof()) throw bad_lexical_cast(); and in dynamic_bitset.hpp, where operator>>(std::basic_istream<CharT, Traits>&, dynamic_bitset<Block, Allocator>&) iterates over the bitset using the size() (which is, of course, 0). I sure could solve my particular problem by providing an explicit template instantiation of the GetValue() function, but I don't want to do this because the code is for a library, and I want to reduce references to other libraries to an absolute minimum. Did I miss something here, or is there a workaround for this problem? Regards, Klaus

1 0