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

RE: [Boost-Users] Graphs: vertex_index vs. edge_index
by Siegel, Matthew 20 May '02

20 May '02

Hi Jeremy, Thanks for the quick response. As you may or may not recall from last week, I am having the issue with MSVC++, where I can not put more than one property per edge or vertex when defining the graph, that limits the size of the template expansion. Because of this, I have to use external property maps that I create myself. Therefore I need an edge index map internal to the graph. Currently my workaround is to keep an unsigned integer in my class, that gets incremented until there's a free "slot" for the edge to go into. This is to allow the increment of the unsigned int to roll at maximum value. Here's the way I think it should work. I understand the space issues that most people would be concerned about, however if a property_map of edge_index_t is explicitly given in the list of edge properties when the graph is defined, then the map should be created automatically for me, and if an index is not provided on the add_edge() function, the index is incremented in the same fashion. If I'm explicitly adding the edge_index map to my graph, then the space is of no concern to me. Just a thought . . .thanks again, Matt -----Original Message----- From: Jeremy Siek [mailto:jsiek@cs.indiana.edu] Sent: Monday, May 20, 2002 12:01 PM To: 'Boost-Users(a)yahoogroups.com' Subject: Re: [Boost-Users] Graphs: vertex_index vs. edge_index Hi Matt, On Mon, 20 May 2002, Siegel, Matthew wrote: > Please correct me if I'm wrong. It seems as if a vertex_index gets > assigned to a vertex as it's added to the graph automatically. > However, this doesn't seem to be the case for the edge_index. > > Is this correct?? That is mostly correct. The vertex_index gets assigned automatically for adjacency_list with VertexList=vecS but not with VertexList=listS. You are right about edge_index. > I forsee the following problem. With the graph implemented using a > std::vector for both the vertex list and the edge list, if a vertex is > removed from the graph the index is corrected for the > vertex_descriptor to the new, appropriate index. However, if an edge Right. All vertices are re-indexed, so that after the removal the remaining indices go from 0 to num_vertices-1 > is removed, the index for the edge_descriptor will not be corrected > with the removal of an element in the vector. Also, this causes > headaches when using an external property map. The edge index is > linked directly to a location in my map. > > Any solutions for these problems?? The reason edge_index is not provided automatically is because it takes up space (whereas the vertex_index is does not), and some users are concerned about space issues. Anyways, one possible solution is to somehow make it possible for you to specify some code to be executed when an edge is added or removed from the graph, and this code would adjust the edge indices, and perhaps the external storage too. But before I get ahead of myself, could you describe how you currently work around this problem, and perhaps how you'd like to see it work? Cheers, Jeremy ---------------------------------------------------------------------- Jeremy Siek http://www.osl.iu.edu/~jsiek/ Ph.D. Student, Indiana Univ. B'ton email: jsiek(a)osl.iu.edu C++ Booster (http://www.boost.org) office phone: (812) 855-3608 ---------------------------------------------------------------------- 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

Graphs: vertex_index vs. edge_index
by Siegel, Matthew 20 May '02

20 May '02

Hi, Please correct me if I'm wrong. It seems as if a vertex_index gets assigned to a vertex as it's added to the graph automatically. However, this doesn't seem to be the case for the edge_index. Is this correct?? I forsee the following problem. With the graph implemented using a std::vector for both the vertex list and the edge list, if a vertex is removed from the graph the index is corrected for the vertex_descriptor to the new, appropriate index. However, if an edge is removed, the index for the edge_descriptor will not be corrected with the removal of an element in the vector. Also, this causes headaches when using an external property map. The edge index is linked directly to a location in my map. Any solutions for these problems?? Thanks, Matt

2 1

newbie question
by the_3_project 20 May '02

20 May '02

How do I compile a program that uses the libregex program? It seems that whenever I do this: $ gcc -o te1 te1.cc -lboost_regex I get: ---- /tmp/cc9r42Gv.o: In function `boost::reg_expression<char, boost::regex_traits<char>, allocator<char> >::fail(unsigned int)': /tmp/cc9r42Gv.o(.boost::reg_expression<char, boost::regex_traits<char>, allocator<char> >::gnu.linkonce.t.fail(unsigned int)+0x4e): undefined reference to `boost::bad_expression::bad_expression(basic_string<char, string_char_traits<char>, __default_alloc_template<true, 0> > const &)' /tmp/cc9r42Gv.o: In function `boost::reg_expression<char, boost::regex_traits<char>, allocator<char> >::compile_set_aux(boost::re_detail::jstack<basic_string<char, string_char_traits<char>, __default_alloc_template<true, 0> >, allocator<char> > &, boost::re_detail::jstack<basic_string<char, string_char_traits<char>, __default_alloc_template<true, 0> >, allocator<char> > &, boost::re_detail::jstack<unsigned int, allocator<char> > &, boost::re_detail::jstack<basic_string<char, string_char_traits<char>, __default_alloc_template<true, 0> >, allocator<char> > &, bool, boost::re_detail::_narrow_type const &)': . . . ---- Here is the source to te1.cc ---- #include <iostream> #include <string> #include <boost/regex.hpp> using namespace boost; int main(int argc, char *argv[]) { cmatch what; string str = "XXyX"; regex expression("(\\S\\S){[Ayw]}(\\S)/"); if(regex_match(str,what,expression)) { cout << "what" << endl; } else { cout << "Shit\n"; } return 0; } -------- What am I doing wrong?

3 2

adjacency_list.cpp
by Louis Lavery 20 May '02

20 May '02

Hello, I've bought the book (BGL) and am trying to compile boost_1_27_0/boost/libs/example/adjacency_list.cpp but... ...MSVC6 spits out the dreaded "fatal error C1001: INTERNAL COMPILER ERROR". Intel 5's first (of many) error is e:/boost_1_27_0/boost/detail/iterator.hpp(204) : error: class "std::list<_Ty, _A>::iterator [with _Ty=boost::list_edge<boost::detail::adj_list_gen<boost::adjacency_list<boost ::vecS, boost::listS, boost::undirectedS, boost::property<boost::vertex_ind ex_t, size_t={unsigned int}, boost::property<boost::vertex_color_t, boost::default_color_type, boost::no_property>>, boost::property<boost::edge_name_t, std::string={std::basic_string<char, std::char_traits<char>, std::allocator<char>>}, boost::no_property>, boost::no_property, boost::listS>, boost::listS, boost::vecS, boost::undirectedS, boost::property<boost::vertex_index_t, size_t={unsigned int}, boost::property<boost::vertex_color_t, boost::default_color_type, boost::n o_property>>, boost::property<boost::edge_name_t, std::string={std::basic_string<char, std::char_traits<char>, std::allocator<char>>}, boost::no_property>, boost::no_property, boost::listS>::config::vertex_descriptor={boost::adjacency_list_traits<boost ::vecS, boost::listS, boost::undirectedS>::vertex_descriptor={boost::ct_if_t<boost::adjacency_list _traits<boost::vecS, boost::listS, boost::undirectedS>::is_rand_a ccess={boost::detail::is_random_access<boost::listS>::type={boost::false_typ e}}, size_t={unsigned int}, boost::adjacency_list_traits<boost::vecS, boost::listS, boost::undirectedS>::vertex_ptr={void *}>::type={boost::adjacency_list_traits<boost::vecS , boost::listS, boost::undirectedS>::vertex_ptr={void *}}}}, boost::property<boost::edge_name_t, std::string={std::basic_string<char, std::char_traits<char>, std::allocator<char>>}, boost::no_property>>, _A=std::allocator<boost::list_edge<bo ost::detail::adj_list_gen<boost::adjacency_list<boost::vecS, boost::listS, boost::undirectedS, boost::property<boost::vertex_index_t, size_t={unsigned int}, boost::property<boost::vertex_color_t, boost::default_color_type, boost::no_property>>, boost::property<boost::edge_name_t, std::string={std::basic_string<char, std::char_traits<char>, std::allocator<char>>}, boost::no_property>, boost::no_property, boost::listS>, boost::listS, boost::vecS, boost::undirectedS, boost::property <boost::vertex_index_t, size_t={unsigned int}, boost::property<boost::vertex_color_t, boost::default_color_type, boost::no_property>>, boost::property<boost::edge_name_t, std::string={std::basic_string<char, std::char_traits<char>, std::allocator<char>>}, boost::no_property>, boost::no_property, boost::listS>::config::vertex_descriptor={boost::adjacency_list_traits<boost ::vecS, boost::listS, boost::undirectedS>::vertex_descriptor={boost::ct_if_t<boost::adjacency_list _traits<boost::vecS, boost::listS, boost::undirectedS>::is_rand_access={boost::detail::is_random_access<boost:: listS>::type={boost::false_type}}, size_t={unsigned int}, boost::adjacency_list_traits<boost::vecS, boost::listS, boost::undirectedS>::vertex_ptr={void *}>::type={boost::adjacency_list_traits<boost::vecS, boost::listS, boost::undirectedS>::vertex_ptr={void *}}}}, boost::property<boost::edge_name_t, std::string={std::basic_string<char, std::char_traits <char>, std::allocator<char>>}, boost::no_property>>>]" has no member "difference_type" typedef typename Iterator::difference_type difference_type; My last hope gcc 2.95.2 comes up with... In file included from e:\boost_1_27_0\boost\detail\numeric_traits.hpp:69, from e:\boost_1_27_0\boost\counting_iterator.hpp:57, from e:\boost_1_27_0\boost\pending\integer_range.hpp:15, from e:\boost_1_27_0\boost\graph\detail\adjacency_list.hpp:34, from e:\boost_1_27_0\boost\graph\adjacency_list.hpp:293, from adjacency_list.cpp:31: e:\boost_1_27_0\boost\cstdint.hpp:163: parse error e:\boost_1_27_0\boost\cstdint.hpp:171: #error defaults not correct; you must hand modify boost/cstdint.hpp I'd like to get it working with MSVC6 or Intel5. Any help much appreciated. Louis.

3 4

BLL for other compilers than GCC and Kai C++
by tslettebo1 19 May '02

19 May '02

Hi all. Have anybody got the Boost Lambda Library to work on other compilers than GCC and Kai C++ (which are mentioned in the documentation)? I have Intel C++ 6.0, which is quite standards compliant, but I get a zillion error messages, just by trying to compile boost/libs/lambda/test/bind_tests_simple.hpp. Trying it with MSVC 6.0 gives an ICE, but that's hardly unexpected. :) Also, are there any plans, on ongoing work, to make it more portable? (I post here, rather than the Boost list, as it's about use of a library) Regards, Terje

1 0

Boost threads compile with Sun WS6U2?
by NYSSEN Peter (BMB) 18 May '02

18 May '02

Hi all, I'm working on a multithreaded application on Sun Solaris 8 with Sun Workshop 6 update 2 installed. Currently my application uses the pthreads library. I've been looking at the Boost.threads library and I would like to use this instead of pthreads. I see that the Sun WS compiler is not in the supported toolsets list and there is no appropriate jam file available. In the Compiler Status: solaris page I see that this compiler is being tested, but no refference to the threads lib. Can anybody point me in the right direction as how I should build this lib in my environment? Is it enough if I try to create a jam file? Has anybody already done this (preferably for release 1_28)? What is the reason this compiler is not in the supported toolsets list? Kind regards, Peter Nyssen **** DISCLAIMER **** "This e-mail and any attachment thereto may contain information which is confidential and/or protected by intellectual property rights and are intended for the sole use of the recipient(s) named above. Any use of the information contained herein (including, but not limited to, total or partial reproduction, communication or distribution in any form) by other persons than the designated recipient(s) is prohibited. If you have received this e-mail in error, please notify the sender either by telephone or by e-mail and delete the material from any computer". Thank you for your cooperation. For further information about Proximus mobile phone services please see our website at http://www.proximus.be or refer to any Proximus agent.

3 2

Graph: Multiple edge properties with MSVC
by Siegel, Matthew 18 May '02

18 May '02

Hi, First off . . . great graph library. I can not get Visual C++ 6 SP5 to compile when I try to put multiple properties on the edges. Seems like a problem with the 3rd parameter that a boost::property can take another boost::property. When the 2nd property is taken out of the edge property list, everything works fine. Thanks for the help! Matt Example: // Vertex property - name as a string typedef boost::property<boost::vertex_name_t, std::string> VertexProperties; // Edge property - name as string, weight as a double typedef boost::property< boost::edge_weight_t, double, boost::property< boost::edge_name_t, std::string> > EdgeProperties; // Graph definition typedef boost::adjacency_list< boost::vecS, // Store out-edges in a std::vector boost::vecS, // Store vertex set in a std::vector boost::directedS, // Graph is directed VertexProperties, // Vertex properties EdgeProperties> Graph; // Edge properties Produces the following compiler error: Compiling... routetable.cpp c:\dev\lib\stlport-4.5.3\stlport\stl\char_traits.h(138) : fatal error C1001: INTERNAL COMPILER ERROR (compiler file 'msc1.cpp', line 1794) Please choose the Technical Support command on the Visual C++ Help menu, or open the Technical Support help file for more information Error executing cl.exe. routing.exe - 1 error(s), 0 warning(s)

3 2

Boost in Visual C++ 6.0
by imi123_1999 17 May '02

17 May '02

I am a newbie so please help me a little. I am using Visual C++ 6.0 in order to compile a program given as an example of threads. I am also copying the program below. But the problem is that the program is giving errors of unresolved externals. Please tell me how to change the settings of VC++ options. Here is the program: #include <boost/thread/thread.hpp> #include <iostream> int count = 0; boost::mutex mutex; void increment_count() { boost::mutex::lock lock(mutex); std::cout << "count = " << ++count << std::endl; } int main(int argc, char* argv[]) { boost::thread_group threads; for (int i = 0; i < 10; ++i) threads.create_thread(&increment_count); threads.join_all(); } The errors i am getting are: C:\Windows\Desktop\threadImI\imi.cpp(9) : error C2039: 'lock' : is not a member of 'mutex' c:\program files\microsoft visual studio\vc98 \include\boost\thread\mutex.hpp(35) : see declaration of 'mutex' C:\Windows\Desktop\threadImI\imi.cpp(9) : error C2065: 'lock' : undeclared identifier C:\Windows\Desktop\threadImI\imi.cpp(9) : error C2146: syntax error : missing ';' before identifier 'lock' C:\Windows\Desktop\threadImI\imi.cpp(19) : warning C4508: 'main' : function should return a value; 'void' return type assumed Error executing cl.exe. Creating browse info file... threadImI.exe - 3 error(s), 3 warning(s)

2 4

RE: [Boost-Users] Boost threads compile with Sun WS6U2?
by schalk_cronje＠nai.com 17 May '02

17 May '02

> I'm working on a multithreaded application on Sun Solaris 8 with Sun > Workshop 6 update 2 installed. > Currently my application uses the pthreads library. > I've been looking at the Boost.threads library and I would like to use > this instead of pthreads. We use it with suncc and build it from a makefile. (We don't actually build any of the boost libs using Jam).

1 0

Call for Volunteers
by William E. Kempf 16 May '02

16 May '02

The Boost-install mail list is up and running (http://lists.boost.org/mailman/listinfo.cgi/boost-install to subscribe), and several people have subscribed. However, traffic is non-existent on the list. I'm concerned that the effort will die if we don't start some momentum. Currently, I have 4 people who volunteered to help with this effort. Hopefully these volunteers are just pressed for time right now and will start contributing shortly, but in any event we need more volunteers. So, I'm calling on people once again. The goal is to create a set of procedures for both the Boost developers and the Boost "installers" that will facilitate the creation of an installation "script" or set of "scripts" for supported platforms. The volunteers will help to define the procedures and to implement the "script(s)" for the target platforms so that Boost users can easily install Boost for use. We need volunteers with a lot of different qualifications. We need someone familiar with every supported platform (I don't mean that a single volunteer needs to know every platform) so that we can insure we define the procedures correctly and can provide an installation solution for every platform. We need people to help us document the procedures and to code the "script(s)". If you're interested in helping with this effort, please join the Boost-install list and let us know how you'd like to contribute. Thanks, Bill Kempf

1 0