April 2007 - Boost - lists.preview.boost.org

Boost topological_sort attempts to advance a dereferencable iterator 20 steps outside the valid range
by Stephen Torri 17 Apr '07

17 Apr '07

I am still trying to understand why boost graph's topological_sort fails for a test program I have made to sort five vertexes. Originally I compiled this program against boost 1.33.1 on a Fedora Core 6 system using gcc-4.1.1. After receiving no response from the boost-users list that could solve this problem I read the Reporting Bug document on the boost website. I followed the CVS instructions and checked out a copy from the HEAD (reported as boost-1.35) and built it. I used the following flags: bjam -sHAVE_ICU=1 -d2 -sTOOLS=gcc -sBUILD=release -sEXPAT_INCLUDE=/usr/include -sEXPAT_LIBPATH=/usr/lib stage Afterwards I installed it and rebuilt the test program. It is still failing inside the topological_sort. The output of the program is below along with the test program source code. Stephen --------- OUTPUT ------------ [storri@localhost infrastructure]$ ./test_dag Added edge: 1 -> 20 Number of edges is 1 Added edge: 1 -> 30 Number of edges is 2 Added edge: 1 -> 40 Number of edges is 3 Added edge: 1 -> 5 Number of edges is 4 Added edge: 20 -> 5 Number of edges is 5 Added edge: 30 -> 5 Number of edges is 6 Added edge: 40 -> 5 Number of edges is 7 /usr/lib/gcc/i386-redhat-linux/4.1.1/../../../../include/c ++/4.1.1/debug/safe_iterator.h:277: error: attempt to advance a dereferenceable (start-of-sequence) iterator 20 steps, which falls outside its valid range. Objects involved in the operation: iterator @ 0x0xbfead0fc { type = N11__gnu_debug14_Safe_iteratorIN9__gnu_cxx17__normal_iteratorIPN5boost18default_color_typeEN10__gnu_norm6vectorIS4_SaIS4_EEEEEN15__gnu_debug_def6vectorIS4_S8_EEEE (mutable iterator); state = dereferenceable (start-of-sequence); references sequence with type `N15__gnu_debug_def6vectorIN5boost18default_color_typeESaIS2_EEE' @ 0x0xbfead0fc } Aborted ----------- CODE --------------- #include <boost/graph/adjacency_list.hpp> #include <boost/shared_ptr.hpp> #include <boost/cstdint.hpp> #include <boost/format.hpp> #include <boost/graph/topological_sort.hpp> #include <iostream> #include <map> class Component; class Graph_Base { public: typedef boost::property< boost::vertex_index_t, boost::uint32_t, boost::property< boost::vertex_name_t, boost::shared_ptr<Component> > > VertexProperty_t; typedef boost::adjacency_list< boost::setS, // OutEdgeList boost::setS, // VertexList boost::directedS, // Directed VertexProperty_t > // VertexProperties Graph_t; typedef boost::adj_list_vertex_property_map<Graph_Base::Graph_t, boost::shared_ptr<Component>, const boost::shared_ptr<Component>&, boost::vertex_name_t> Property_Map_const_t; }; class Component { public: typedef boost::shared_ptr<Component> ptr_t; Component ( boost::uint32_t id ) : m_id ( id ) {} ~Component () { m_source_list.clear(); } // Now part of Component_Data void set_Data_Source ( uint32_t id ) { if ( ! m_source_list.insert ( std::make_pair ( id, false ) ).second ) { std::cerr << "WARNING: Attempted to add a duplicate component id."; std::cerr << std::endl; std::cerr << "Ignoring component id." << std::endl; } } std::map<boost::uint32_t, bool>::const_iterator get_Source_List_Begin() const { return m_source_list.begin(); } std::map<boost::uint32_t, bool>::const_iterator get_Source_List_End() const { return m_source_list.end(); } boost::uint32_t get_ID (void) const { return m_id; } private: boost::uint32_t m_id; std::map<boost::uint32_t, bool> m_source_list; }; class Component_Graph { public: typedef boost::graph_traits<Graph_Base::Graph_t>::vertex_descriptor Vertex_t; typedef boost::graph_traits<Graph_Base::Graph_t>::edge_descriptor Edge_t; typedef std::map<uint32_t, Vertex_t> IdVertexMap_t; Component_Graph () {} void add_Component ( Component::ptr_t obj_ptr ) { bool no_duplicate; IdVertexMap_t::iterator pos; Vertex_t node; if ( obj_ptr.get() == 0 ) { std::cerr << boost::format("Exception throw in %s at line %d") % __FILE__ % __LINE__ << std::endl; return; } boost::tie (pos, no_duplicate) = m_index.insert (std::make_pair(obj_ptr->get_ID(), node)); if ( no_duplicate ) { // Insertion was successful // Make a new vertex and return handle to 'node' node = add_vertex(m_graph); // Get list of Components boost::property_map<Graph_Base::Graph_t, boost::vertex_name_t>::type clist = boost::get(boost::vertex_name, m_graph); // Assign component to vertex position clist[node] = obj_ptr; // Add vertex handle to map position pos->second = node; // Get list of indexes boost::property_map<Graph_Base::Graph_t, boost::vertex_index_t>::type ilist = get(boost::vertex_index, m_graph); ilist[node] = obj_ptr->get_ID(); for ( std::map<boost::uint32_t,bool>::const_iterator cpos = obj_ptr->get_Source_List_Begin(); cpos != obj_ptr->get_Source_List_End(); ++cpos ) { this->add_Child ( (*cpos).first, obj_ptr ); } } else { std::cerr << "ERROR: Duplicate source found. Skipping Component" << std::endl; return; } } void add_Child ( uint32_t parent_id, Component::ptr_t obj_ptr ) { IdVertexMap_t::iterator pos; IdVertexMap_t::iterator cpos; if ( obj_ptr.get() == 0 ) { std::cerr << boost::format("Exception throw in %s at line %d") % __FILE__ % __LINE__ << std::endl; return; } pos = m_index.find (parent_id); cpos = m_index.find ( obj_ptr->get_ID() ); if (pos == m_index.end()) { std::cerr << "ERROR: Cannot find parent. Skipping adding component #" << obj_ptr->get_ID() << std::endl; return; } // Add edge from parent to child Edge_t link; if (! (add_edge ( pos->second, cpos->second, m_graph).second)) { std::cerr << "ERROR: Duplicate edge exists from " << parent_id << " to " << obj_ptr->get_ID() << std::endl; } else { std::cout << boost::format("Added edge: %d -> %d") % parent_id % obj_ptr->get_ID() << std::endl; std::cout << "Number of edges is " << num_edges ( m_graph ) << std::endl; } } Graph_Base::Graph_t const& get_Graph () const { return m_graph; } Component::ptr_t get_Component (const Vertex_t& node) const { Graph_Base::Property_Map_const_t clist = get(boost::vertex_name, m_graph); return clist[node]; } private: Graph_Base::Graph_t m_graph; IdVertexMap_t m_index; }; int main (int, char**) { // CREATE Component_Graph Component_Graph m_graph; // Create Components Component::ptr_t node_one ( new Component ( 1 ) ); Component::ptr_t node_twenty ( new Component ( 20 ) ); node_twenty->set_Data_Source ( 1 ); Component::ptr_t node_thirty ( new Component ( 30 ) ); node_thirty->set_Data_Source ( 1 ); Component::ptr_t node_forty ( new Component ( 40 ) ); node_forty->set_Data_Source ( 1 ); Component::ptr_t node_five ( new Component ( 5 ) ); node_five->set_Data_Source ( 1 ); node_five->set_Data_Source ( 20 ); node_five->set_Data_Source ( 30 ); node_five->set_Data_Source ( 40 ); // Add Components to Graph m_graph.add_Component ( node_one ); m_graph.add_Component ( node_twenty ); m_graph.add_Component ( node_thirty ); m_graph.add_Component ( node_forty ); m_graph.add_Component ( node_five ); // Call topological_sort on graph typedef std::vector<Component_Graph::Vertex_t> m_container; m_container comp_list; boost::topological_sort ( m_graph.get_Graph(), std::back_inserter ( comp_list ) ); std::cout << "A topological ordering: "; for ( m_container::reverse_iterator ii = comp_list.rbegin(); ii != comp_list.rend(); ++ii ) { Component::ptr_t comp_ptr = m_graph.get_Component ( *ii ); std::cout << comp_ptr->get_ID() << " "; } std::cout << std::endl; return 0; }

1 0

[Report] 3 regressions on RC_1_34_0 (2007-04-17)
by Douglas Gregor 17 Apr '07

17 Apr '07

Boost Regression test failures Report time: 2007-04-17T04:05:39Z This report lists all regression test failures on release platforms. Detailed report: http://engineering.meta-comm.com/boost-regression/CVS-RC_1_34_0/developer/i… 3 failures in 1 libraries utility (3) |utility| none_test: msvc-6.5 msvc-6.5_stlport4 msvc-7.0

1 0

range, multi-dimensional
by Neal Becker 17 Apr '07

17 Apr '07

I just posted something on ublas, but I think it may be of more general interest. I have been using boost::range heavily, and find it useful for generic interfaces. I think it is also useful to have a multi-dimensional extension to range. In particular, a 2-d extension would help in creating algorithms that can accept a variety of 2-d structures. For example, range has size, begin, end. range2d would have size1, size2, begin1, begin2. Other useful members might include row, col.

7 14

[Report] 2 regressions on RC_1_34_0 (2007-04-14)
by Douglas Gregor 17 Apr '07

17 Apr '07

Boost Regression test failures Report time: 2007-04-13T23:33:32Z This report lists all regression test failures on release platforms. Detailed report: http://engineering.meta-comm.com/boost-regression/CVS-RC_1_34_0/developer/i… 2 failures in 2 libraries serialization (1) utility (1) |serialization| test_split_binary_archive_dll: msvc-8.0 |utility| none_test: borland-5.8.2

6 7

[Report] 0 regressions on RC_1_34_0 (2007-04-13)
by Douglas Gregor 17 Apr '07

17 Apr '07

Boost Regression test failures Report time: 2007-04-13T06:40:52Z This report lists all regression test failures on release platforms. Detailed report: http://engineering.meta-comm.com/boost-regression/CVS-RC_1_34_0/developer/i… No failures! Yay!

5 4

[Boost-bugs] [ boost-Bugs-1702032 ] [1.33.1]Paths not equivalent on Windows when casing differs
by SourceForge.net 17 Apr '07

17 Apr '07

Bugs item #1702032, was opened at 2007-04-17 01:19 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1702032&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: filesystem Group: None Status: Open Resolution: None Priority: 5 Private: Yes Submitted By: Nobody/Anonymous (nobody) Assigned to: Beman Dawes (beman_dawes) Summary: [1.33.1]Paths not equivalent on Windows when casing differs Initial Comment: If the casing differs of the native paths stored in boost::filesystem::path objects. Casing should not matter on Windows. The code snippet below demonstrates the problem: path p1( "C:\\Windows", native ); path p2( "C:\\Windows", native ); path p3( "c:\\Windows", native ); set<path> paths; paths.insert( p1 ); // This one is fine. If the casing of the // native paths is the same, it is found // in the set assert( paths.end() != paths.find( p2 ) ); // This one fires, although it shouldn't // on Windows. The paths are equivalent assert( paths.end() != paths.find( p3 ) ); ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1702032&group_… ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Boost-bugs mailing list Boost-bugs(a)lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/boost-bugs

1 0

[Boost-bugs] [ boost-Bugs-1701992 ] [program_options] improper guess
by SourceForge.net 17 Apr '07

17 Apr '07

Bugs item #1701992, was opened at 2007-04-17 00:11 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1701992&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Alexis Wilke (alexis_wilke) Assigned to: Nobody/Anonymous (nobody) Summary: [program_options] improper guess Initial Comment: I found a problem in the program_options when using the allow_guessing style (which is the default). The option_description::match() function (boost_1_33_1/libs/program_options/src/options_description.cpp) checks option names approximatively only if the allow_guessing is turned on. The match() function needs to return three values: definitively not this option, definitively this option or maybe this option. In the first case, you try the next option (see options_description::find_nothrow() in the same file). In the second case, you return with that very option. In the third case, you try the next options. If at the very end, you did not find a definitive option and you found more than 2 matching approximetly, then you generate the error (more than one approx. matched!). And if nothing matched, you generate that other error. This is only partially fixed in Debian and I'm attaching that fix below. The problem in the Debian fix is that if the valid full definitive option is to be found AFTER two other partial options, it also breaks (duh! am I the only one to understand this logic?! it seems so simple!) I did not understand the 1.34.0 yet and thus I did not see the fix there... ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1701992&group_… ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Boost-bugs mailing list Boost-bugs(a)lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/boost-bugs

1 0

boost::freetype - extension for boost::gil
by Tom Brinkman 17 Apr '07

17 Apr '07

>> I'm sorry; I didn't want to suggest I gave an exhaustive enumeration >> of the kinds of processing that are necessary. I hadn't realised you >> provide an interface for retrieving ligatures. >> You seem to suggest that your framework will work with other languages >> than English (correct me if I'm wrong). Arabic requires marks to be >> positioned and glyphs to be replaced based on their context. As I see it, the glyph rendering should be done seperately from the contextual repositioning of glyphs based upon their context. So for "arabic" language support, one would have to arrange the glyphs prior to passing it to their renderer. I agree, this would require you to pass the entire string into the processing function. Example: //step 1: convert string to glyph vector std::string str = "abcdefg"; std::vector<glyph> glyphs; boost::make_arabic_glyphs(str, std::back_inserter(glyphs)); >> Having done some work in the area I know that rendering text is fairly >> complicated. What is your reason for not hiding this complexity? What >> are the gains of accessing individual glyphs? I'll answer your question with a motivating example. This example shows how to find the last glyph to fit into a view and calculate its total width (in pixels). Once I have the "fitted" width, its a simple matter to position the glyph array onto a "x-y" axis. I think that this code is pretty basic and most any c++/boost developer would be very comfterble with this. //step 1: make metric vector std::vector<FT_Glyph_Metrics> metrics; std::transform(glyphs.begin(),glyphs.end(), std::back_inserter(metrics), boost::freetype::make_metric()); //step 2: find last fitted glyph std::vector<FT_Glyph_Metrics>::iterator it = std::find_if(metrics.begin(),metrics.end(),kerning.begin(), boost::freetype::find_last_fitted_glyph(view.width())); //step 3: find the width int distance = (int)std::distance(it,metrics.end()); int width = std::for_each(metrics.begin(), metrics.end()-distance, kerning.begin(), boost::freetype::make_width());

1 0

[automata] request for comments
by Cromwell Enage 16 Apr '07

16 Apr '07

With all but one of the tutorials mostly finished, I should go ahead and announce the existence of my automata library, a dynamic state machine framework which you can either use to construct DFAs and DPDAs or extend to build your own types of automata. If you haven't done so already, you can download it from http://boost-consulting.com/vault/ (automata.zip is in the Data Structures directory). If you have an older version, here are some changes: * The automaton<> front-end class template used to require only one template parameter: a Transition Function. Now, it also needs a Unary Metafunction Class that positionally comes first and whose return type is an Automaton Base. The decoupling was done to allow for cases like a user-defined type instead of dfa_generator working together with a matrix_transition_function<>, etc. * Visitors/Traits are now Observers. * Input-Check Visitors are now Input Validators. * Renamed *_ostream_policy<> class templates to *_insertion_policy<> class templates, to allow for the use of a Boost.Serialization archive as an Insertion Target instead of an output stream. * Renamed *_check_policy<> class templates to *_validation_policy<> class templates. * builder_from_graph<> now lets you attach Observer Builders to it. * Added an option to enable undo functionality. * Removed unnecessary type definitions. Concept and reference documentation still need tons of work, but I think the code structure is stable enough for preliminary testing. Feedback is much appreciated, especially in the following areas: * Tutorials. Are they comprehensive enough? What other potential use cases deserve attention? * Interface. Is it too restrictive? Is it too big? To all of you who've been testing this library at its various stages, thank you for your continued interest! Cromwell D. Enage __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

1 0

[Boost-bugs] [ boost-Bugs-1701807 ] exists() throws exeption on existing but locked file on wind
by SourceForge.net 16 Apr '07

16 Apr '07

Bugs item #1701807, was opened at 2007-04-16 15:52 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1701807&group_… Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: filesystem Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Hartmut Kaiser (hkaiser) Assigned to: Beman Dawes (beman_dawes) Summary: exists() throws exeption on existing but locked file on wind Initial Comment: boost::filesystem::exists(somepath) throws an exception from when 'somepath' exists but is locked by some other process (at least on Windows). A simple way to reproduce the effect is: #include <iostream> #include <boost/filesystem.hpp> int main() { boost::filesystem::path p("C:\\pagefile.sys"); try { boost::filesystem::exists(p); } catch (...) { std::cerr << "Ooops..." << std::endl; } return 0; } Throwing an exception doesn't seem to be the right thing here, since the file actually exists, so I'ld expect to get a 'true' back from exists(). ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1701807&group_… ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Boost-bugs mailing list Boost-bugs(a)lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/boost-bugs

1 0