- Boost-users - lists.boost.org

Bug Sprint Report for Thursday, June 4th
by Marshall Clow 05 Jun '09

05 Jun '09

The boost bug sprint started Friday, May 29th, and runs through Sunday June 7th. Thursday was a very good day - let's keep it up! The bug sprint will be over Sunday night. Status <https://svn.boost.org/trac/boost/report/1> 7:00 AM PDT 5/29: 797 open tickets. 7:45 AM PDT 5/30: 772 open tickets. 6:00 AM PDT 5/31: 761 open tickets. 6:00 AM PDT 6/2: 747 open tickets. 6:00 AM PDT 6/3: 748 open tickets. 6:00 AM PDT 6/4: 739 open tickets. 8:00 AM PDT 6/5: 730 open tickets. Activity: 5/29: 6 new tickets, 77 modifications, 25 tickets closed 5/30: 2 new tickets, 42 modifications, 17 tickets closed 5/31: 2 new tickets, 27 modifications, 8 tickets closed 6/1: 2 new tickets, 42 modifications, 17 tickets closed 6/2: 5 new tickets, 18 modifications, 5 tickets closed. 6/3: 4 new tickets, 19 modifications, 2 tickets closed. 6/4: 3 new tickets, 42 modifications, 24 tickets closed. [ Numbers may not add up because I am counting via two different methods, using data that is time-stamped in two different time zones. Also, tickets can be closed and reopened. ] What can you do? See <https://svn.boost.org/trac/boost/wiki/BugSprintJune2009> for details. I've added four new reports to the trac: Open ticket count by owner: <https://svn.boost.org/trac/boost/report/18> Open ticket count by component: <https://svn.boost.org/trac/boost/report/19> Open ticket count by ticket type: <https://svn.boost.org/trac/boost/report/20> Open ticket count by milestone: <https://svn.boost.org/trac/boost/report/21> There's been a lot of discussion on the list about tickets; and there are about a dozen patches that have been tested and are ready to be applied. If you're looking to help, you can look at some of the older bugs, say, the ones for Boost 1.36, and determine if they still apply. If they do, note that in the ticket. There are 35 tickets that are still open that are marked as "to be fixed in Boost 1.36" See <https://svn.boost.org/trac/boost/report/3> for a list of open tickets by milestone. --------- Sign-up sheet -------- The following people have committed to participate in the bug sprint. Please add you name here, along with what you can do. 1) Marshall Clow (coordination, Mac OS X characterization) 2) Beman Dawes (filesystem, system, timer) 3) Doug Gregor (function, graph, logic, mpi, numeric, property_map, signals) 4) "joel falcou" <joel.falcou(a)u-psud.fr> ( linux gcc testing/VC 2008 testing, patching) 5) Michael Caisse <boost(a)objectmodelingdesigns.com> (bug fixing) 6) Jeremiah Willcock <jewillco(a)osl.iu.edu> (will test and commit graph patches) 7) David Abrahams (iterator, python) 8) Sohail Somani (patches for g++ 4.2.4) 9) Steven Watanabe <watanabesj(a)gmail.com> (bug fixes, como) 10) vicente.botet" <vicente.botet(a)wanadoo.fr> 11) Stjepan Rajko <stjepan.rajko(a)gmail.com> (fusion) 12) Daniel Wallin (parameter, ...) -- -- Marshall Marshall Clow Idio Software <mailto:marshall@idio.com> It is by caffeine alone I set my mind in motion. It is by the beans of Java that thoughts acquire speed, the hands acquire shaking, the shaking becomes a warning. It is by caffeine alone I set my mind in motion.

1 0

BGL: how to count and get the connected components in the spanning_tree
by arm2arm 05 Jun '09

05 Jun '09

Hello, I am trying to get connected components in the MST. I have some undirected graph: ----------- BEGIN SNIP ------------------------ typedef adjacency_list < vecS, vecS, undirectedS, no_property, property < edge_weight_t, MyFloat > > Graph; typedef property_map<Graph, edge_weight_t>::type EdgeWeightMap; Graph graphFOF; ... Fill the graph.. //Get the MST property_map < Graph, edge_weight_t >::type weight = get(edge_weight, graphFOF); std::vector < Edge > spanning_tree; std::cout << "Building MST:." ; kruskal_minimum_spanning_tree(graphFOF, std::back_inserter(spanning_tree)); std::cout << " .. done " << std::endl; // // Here is problem: How to use MST edges for filtering the graphFOF?? positive_edge_weight<EdgeWeightMap> filter(meanW, get(edge_weight, graphFOF)); filtered_graph<Graph, positive_edge_weight<EdgeWeightMap> > fg(graphFOF, filter); std::vector<int> component(num_vertices(fg)); int num = connected_components(fg, &component[0]); ------------END SNIP--------------------------- How to count and get the connected components in the spanning_tree? Thank you beforehand. Arman. -- View this message in context: http://www.nabble.com/BGL%3A-how-to-count-and-get-the-connected-components-… Sent from the Boost - Users mailing list archive at Nabble.com.

1 0

Re: [Boost-users] About "transitive_closure" method of Boost graph.
by Amanullah Yasin 05 Jun '09

05 Jun '09

sorry for disturbing, i think you mean test example given on http://www.boost.org/doc/libs/1_39_0/libs/graph/example/transitive_closure.… Yes i tried this example but it is not using this method. void transitive_closure(const Graph& g, GraphTC& tc, G_to_TC_VertexMap g_to_tc_map, VertexIndexMap index_map) Thanks again Amanullah YASIN Master2 (ECD) Ecole Polytechnique Université de Nantes, France. Mob# 06 47 62 89 96 Fix # 02 72 00 26 45 amanyasin(a)gmail.com On Fri, Jun 5, 2009 at 10:57 AM, Vladimir Prus <vladimir(a)codesourcery.com>wrote: > On Friday 05 June 2009 12:52:37 you wrote: > > Thanks Vladimir Prus, for your reply. Please guide me how i can use this > > method > > Please reply to the list, not personally. > > > > > > > void transitive_closure(const Graph& g, GraphTC& tc, > > G_to_TC_VertexMap g_to_tc_map, VertexIndexMap > index_map) > > > > > > *specially how i can declare and use * > > > > * "G_to_TC_VertexMap" & "* > > > > *VertexIndexMap"* > > I have suggested you look at the testcase -- did you do that? > > - Volodya >

1 0

[multi-index] Container not safe against reentrant clear() / erase() ?
by Sebastian Theophil 05 Jun '09

05 Jun '09

Hi, I have encounted a scenario that triggers an access violation when removing reference counted objects from a container. Assume I have a (for simplicity's sake, global) multi_index_container g_multiindex containing reference counted objects of type CObject: boost::multi_index_container< ordered_unique< shared_ptr<CObject> > > g_multiindex; In general, an object of type CObject may be destroyed at any time and the destructor makes sure to remove any reference held in g_multiindex: class CObject { ~CObject() { // if object is explicitly destroyed, remove from container g_multiindex.erase( this ); } }; Since ownership of CObject is shared, g_multiindex may hold the last reference to a CObject instance. g_multiindex.clear then calls - delete_all_nodes() - delete_all_nodes(root()) - this->final_delete_node_(static_cast<final_node_type*>(x)); - which first calls the destructor on x (i.e. it calls ~CObject() on the root() node) At this point, the multi_index_container is not in a valid state. The root() node is being deleted but the container still contains the root() node and points to it, empty() would certainly still return false. Calling g_multiindex.erase with a compatible key crashes in my case. In general, it seems to me that it would be safer to implement clear() by doing something like: void clear() { multi_index_container cpy; swap(cpy, *this); cpy.clear(); } This would remove all elements from the container before actually deleting them. I'm unsure though how the same behavior could be achieved when erasing a single element. The C++ standard does not seem to define how clear() and erase() should behave in this regard, but I cannot image that my application scenario is so unusual that this problem has not been solved a million times. So I'm curious if the behavior of multi_index_containers is by design and maybe someone could be so kind to enlighten me why this is so. Is there any better way to achieve what I'm trying to do? Thank you very much, Sebastian -- Sebastian Theophil . stheophil(a)think-cell.com Software Engineer think-cell Software GmbH . Invalidenstr. 34 . 10115 Berlin, Germany http://www.think-cell.com . phone +49-30-666473-10 . toll-free (US) +1-800-891-8091 Directors: Dr. Markus Hannebauer, Dr. Arno Schoedl . Amtsgericht Berlin-Charlottenburg, HRB 85229

4 8

About "transitive_closure" method of Boost graph.
by Amanullah Yasin 05 Jun '09

05 Jun '09

Bonjour Dear I want to check whether there is a directed path between two given nodes (A and C) using Boost graph. I also want to access intermediate edges to examine PDAG properties. I test void transitive_closure(const Graph& g, GraphTC& tc) method. But i want to check between two given nodes. Now i want to ask you - void transitive_closure(const Graph& g, GraphTC& tc, G_to_TC_VertexMap g_to_tc_map, VertexIndexMap index_map) In the above method, how i can declare G_to_TC_VertexMap g_to_tc_map, VertexIndexMap index_map? http://www.boost.org/doc/libs/1_39_0/libs/graph/doc/transitive_closure.html - How i can examine the intermediate edges? or is there any other easy way? I will be thankful to you Best Regards Amanullah YASIN Master2 (ECD) Ecole Polytechnique Université de Nantes, France. Mob# 06 47 62 89 96 Fix # 02 72 00 26 45 amanyasin(a)gmail.com

2 2

Refactoring template functions. Needing help with compile time if statements.
by Ryan McConnehey 05 Jun '09

05 Jun '09

I'm trying to refactor my code to remove some code duplication. Currently I have a template that takes a pointer to store information, a map with the information and a text string. I've specialized the function to account for some fundamental types. As these cases have increased I notice the only difference between the function is how I'm accessing the iterator for the map. template<> void Customize::addVariableType(TiXmlElement *const xml, std::map<string, bool> const& map, string const& mapType) const { std::map<string, bool>::const_iterator itor; for (itor = map.begin(); itor != map.end(); ++itor) { itor->second; } } template<typename T> void Customize::addVariableType(TiXmlElement *const xml, std::map<string, T> const& map, string const& mapType) const { std::map<string, T>::const_iterator itor; for (itor = map.begin(); itor != map.end(); ++itor) { itor->second.base_type_value(); } } I'm using stlSoft's true_typedef for some custom types. STLSOFT_GEN_OPAQUE(fundamental_type); typedef stlsoft::true_typedef<unsigned short, fundamental_type> uint16_type; Most of the maps are comprised of these custom types. That's why the generic template accesses the underlying value of the iterator's second value. Also this accounts for the specialization for fundamental types. Since both the template and the specialization template have the same functionality but access the second map value differently I'm trying to refactor to have another function retrieve iterator value. Below is what I've tried so far. template<> bool accessValue(bool const& value) const { return value; } //V is the underlining type of W. If W was uint16_type then V should be unsigned short. template<typename V, typename W> typename V accessValue(typename W const& value) const { return value.base_type_value(); } I understand that I've moved the difficulty out of addVariableType into accessValue. Unfortunately, accessValue gives an error that typename V cannot be deduced. Also for each fundamental type I have to add an overload to accessValue. This turns into the same problem I'm facing with addVariableType but with less code to copy. Is there a way to have a template function that has an if statement to choice how to access the value? I've tried to use BOOST_TYPEOF_TPL and boost::is_class but was unsuccessful in making it work. Any help would be really helpful. Ryan

1 0

Box-Muller transform Gaussian Random Number Generator
by Enrique Fernández 05 Jun '09

05 Jun '09

Hello, Boost random library implements GRNG (Gaussian Random Number Generator) with Box-Muller transform (code here<http://www.boost.org/doc/libs/1_39_0/boost/random/normal_distribution.hpp>). I want to discuss about this line of code: _cached_rho = sqrt(-result_type(2) * log(result_type(1)-_r2)); I don't know why log is applied to 1-r2, instead of r2 directly, yielding _cached_rho = sqrt(-result_type(2) * log(_r2)); According with [1<http://www.cse.cuhk.edu.hk/%7Ephwl/mt/public/archives/papers/grng_acmcs07.p…>, 2 <http://mathworld.wolfram.com/Box-MullerTransformation.html>, 3<http://en.wikipedia.org/wiki/Box-Muller_transformation>] the uniform random variable r2 is always used directly (never 1-r2). I'd like to know why 1-r2 is used. I've tried to compare with other scientific libraries like GSL<http://www.gnu.org/software/gsl/manual/html_node/The-Gaussian-Distribution.…>, Scythestat <http://scythe.wustl.edu/api/rng_8h-source.html> and LTILib<http://ltilib.sourceforge.net/doc/html/index.shtml>, but they use the Box-Muller polar form (or Marsaglia polar method) or even Ziggurat algorithm, which are supposed to be more efficient. Why isn't polar form implemented? I'd like to know why 1-r2 is used anyway. Regards, Enrique

2 1

What will be in v1.40?
by Robert Dailey 05 Jun '09

05 Jun '09

Is there a document online that covers new features that will be introduced in Boost v1.40? If not, can someone provide a brief overview? Thanks.

2 1

Re: [Boost-users] Boost-users Digest, Vol 2016, Issue 2
by Tron Thomas 04 Jun '09

04 Jun '09

> Date: Thu, 4 Jun 2009 09:44:34 +0100 > From: "John Maddock" <john(a)johnmaddock.co.uk> > Subject: Re: [Boost-users] Compile error using boost's tr1::bind > To: <boost-users(a)lists.boost.org> > Message-ID: <059F08EC9A4049B194EBF6D5CA604C4C@acerlaptop> > Content-Type: text/plain; format=flowed; charset="Windows-1252"; > reply-type=original > > >> What is needed for the code to compile and run properly? >> > > Which Boost version and is Boost installed in /usr/include or somewhere > similar? > > Basically this is a known problem for some Boost versions when Boost is > installed in your *system search path* - the issue is that the Boost headers > are in a location that is searched *after* the libstdc++ headers and that > breaks the header-forwarding mechanism used by Boost.TR1. > > The solution is probably to upgrade to the latest Boost release, or if all > else fails, the SVN Trunk. > > HTH, John. > I'm not sure which version of boost this is. It would be whatever the most current version of boost is for Fedora 10, as I update my system on a regular basis. There is a /usr/include/boost directory on my system. I would assume that is the version of boost being used by the compiler.

2 1

unit tests: auto tests, missing main, redef'ed init_unit_test_suite
by Hicham Mouline 04 Jun '09

04 Jun '09

g++3.4 - boost1.37 Hello, I have n translation units .cpp and m auto test suites spread over the n files. For e.g. file ni.cpp-------------------- #include <boost/test/unit_test.hpp> BOOST_AUTO_TEST_SUITE(suite1) BOOST_AUTO_TEST_CASE(case1) BOOST_AUTO_TEST_SUITE_END() file nj.cpp-------------------- #include <boost/test/unit_test.hpp> BOOST_AUTO_TEST_SUITE(suite1) BOOST_AUTO_TEST_CASE(case2) BOOST_AUTO_TEST_CASE(case3) BOOST_AUTO_TEST_SUITE_END() file nk.cpp-------------------- #include <boost/test/unit_test.hpp> BOOST_AUTO_TEST_SUITE(suite2) BOOST_AUTO_TEST_CASE(case1) BOOST_AUTO_TEST_CASE(case2) BOOST_AUTO_TEST_SUITE_END() The .cpp files are compiled with -DBOOST_TEST_MAIN= AFAIU, I am then not required to provide the "init_unit_test" function nor the "main" function. I then try to link an executable: g++ -m64 -Wall -Wextra -Wsign-compare -Werror -Wno-unused -std=c++98 -pedantic -Wno-long-long -pipe -g -lboost_unit_test_framework-gcc34-mt-d-1_37 -o executable *.o but i get this: nk.o: In function `init_unit_test_suite(int, char**)': ...boost/test/unit_test_suite.hpp:188: multiple definition of `init_unit_test_suite(int, char**)' ni.o: ...boost/test/unit_test_suite.hpp:188: first defined here for each object file and /usr/lib/gcc/x86_64-redhat-linux/3.4.6/../../../../lib64/crt1.o(.text+0x21): In function `_start': : undefined reference to `main' Regards,

3 4