- Boost-users - lists.boost.org

newbie problem with Boost.Serialization
by John Pye 26 Apr '07

26 Apr '07

Hi all, I have some trouble with the Boost.Serialization library. I have compiled the 'demo.cpp' example, and now I am trying to distill it a little, so that I can understand how to apply it to my own polymorphic class hierarchy. However I keep seeing the same error message, and can't work out what it means. In /usr/include/boost/archive/detail/oserializer.hpp:567 there are some comments about the error message I'm seeing but I don't understand what those comments mean. I'm using Boost 1.33.1-7ubuntu1 on Ubuntu 6.10. Any help would be much appreciated! Cheers JP Here is the error message: --------------8<--------------- jpye@john:~/boosttest$ scons scons: Reading SConscript files ... scons: done reading SConscript files. scons: Building targets ... g++ -o ser1.o -c ser1.cpp /usr/include/boost/archive/detail/oserializer.hpp: In function 'void boost::archive::save(Archive&, T&) [with Archive = boost::archive::text_oarchive, T = dd]': /usr/include/boost/archive/basic_text_oarchive.hpp:78: instantiated from 'void boost::archive::basic_text_oarchive<Archive>::save_override(T&, int) [with T = dd, Archive = boost::archive::text_oarchive]' /usr/include/boost/archive/detail/interface_oarchive.hpp:78: instantiated from 'Archive& boost::archive::detail::interface_oarchive<Archive>::operator<<(T&) [with T = dd, Archive = boost::archive::text_oarchive]' ser1.cpp:43: instantiated from here /usr/include/boost/archive/detail/oserializer.hpp:567: error: invalid application of 'sizeof' to incomplete type 'boost::STATIC_ASSERTION_FAILURE<false>' scons: *** [ser1.o] Error 1 scons: building terminated because of errors. jpye@john:~/boosttest$ -------------------------------------- Here is the code that was being compiled: -----------------8<--------------- jpye@john:~/boosttest$ cat ser1.cpp #include <iomanip> #include <iostream> #include <fstream> #include <string> #include <boost/archive/tmpdir.hpp> #include <boost/archive/text_iarchive.hpp> #include <boost/archive/text_oarchive.hpp> #include <boost/serialization/base_object.hpp> #include <boost/serialization/utility.hpp> #include <boost/serialization/list.hpp> class dd{ friend std::ostream&operator<<(std::ostream&,const dd&); friend class boost::serialization::access; double x,y; template<class Archive> void serialize(Archive & ar, const unsigned int /* file_version */){ ar & y & x; } public: // every serializable class needs a constructor dd(){}; dd(double x, double y) : x(x), y(y) {} }; std::ostream & operator<<(std::ostream &os, const dd &d) { return os << d.x << ' ' << d.y; } int main(int argc, char *argv[]){ dd D(10,20); int x = 5; std::cerr << D << std::endl; std::ofstream ofs("test.txt"); boost::archive::text_oarchive oa(ofs); oa << x; oa << D; return 0; } -------------------------------------- And, just for completeness, my build file: --------------8<---------------- jpye@ascendserver:~/boosttest$ cat SConstruct Program("test",["ser1.cpp"], LIBS=['boost_serialization']) ------------------------------------

2 3

Re: [Boost-users] [Serialization, XML] - how to get info "where am i?" in serialization helper (encoding is not recursive)?
by Frank Bergemann 26 Apr '07

26 Apr '07

Hi Gary, you made my day - many thanks! Seems we have done similar things :-) Just a final question: Could you tell me, what these defines are about? I didn't adopt them and the modules seems to be ok without. > > #define BOOST_ARCHIVE_CUSTOM_IARCHIVE_TYPES datastore::ConfigIArchive > #define BOOST_ARCHIVE_CUSTOM_OARCHIVE_TYPES datastore::ConfigOArchive > Cheers! \Frank _______________________________________________________________ SMS schreiben mit WEB.DE FreeMail - einfach, schnell und kostenguenstig. Jetzt gleich testen! http://f.web.de/?mc=021192

4 5

BOOST_CLASS_EXPORT conflicts with xml_iarchive.hpp
by Oleg V. Zhylin 26 Apr '07

26 Apr '07

Hello All, Could someone please help me to understand why code below fails to compile with STATIC_ASSERTION_FAILURE in basic_xml_iarchive.hpp // If your program fails to compile here, its most likely due to // not specifying an nvp wrapper around the variable to // be serialized. BOOST_STATIC_ASSERT(0 == sizeof(T)); Code does compile if BOOST_CLASS_EXPORT line is commeted out. #include <boost/archive/xml_iarchive.hpp> class A { friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int version) { ar & i; } int i; }; #include <boost/serialization/export.hpp> BOOST_CLASS_EXPORT(A) int main() { } WBR Oleg V. Zhylin ovz(a)yahoo.com __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

3 5

Filesystem - Recurse directories function?
by Jeff 25 Apr '07

25 Apr '07

I am wondering if someone would possibly post a recursive function that prints out all files and sub-directories. I've never created a recursive function before and I'm having problems getting all files and directories to print. Thank You

3 2

Boost topological_sort attempts to advance a dereferencable iterator 20 steps outside the valid range
by Stephen Torri 25 Apr '07

25 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

math toolkit review
by Arnaldur Gylfason 25 Apr '07

25 Apr '07

Hi, This is my contribution to the boost::math review. This is a partial review of components 1) and 2) mentioned in the formal review announcement, i.e. 1) the statistical distributions and 2) special functions. ? What is your evaluation of the design? Very good in my opinion. I think using global functions is very appropriate in libraries such as this one. ? What is your evaluation of the implementation? I did not look at the implementation but I compared it with another implementation and it compares favorably. More on this below. ? What is your evaluation of the documentation? Very clear. I only browsed it, but that was enough to get the feel for the library and start using it right away. ? What is your evaluation of the potential usefulness of the library? Very useful in applications that deal with anything related to statistics or probability. ? Did you try to use the library? With what compiler? Did you have any problems? Yes I used it on a Linux box using g++ 4.0.2. No problems, except I was using version 1.33.0 of boost so I had to change all use of tuple to be in namespace tuples. Took a short time though. ? How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? I guess a quick reading plus I wrote 2 small test programs that compare accuracy and speed with implementations from the R-project. ? Are you knowledgeable about the problem domain? I have been working in a development group in a Statistics department of a genetics company for several years so I guess I'm fairly knowledgeable. ? Do you think the library should be accepted as a Boost library? Yes, definitely. The test programs are compareAccuracy.cc and compareSpeed.cc. The programs were compiled with g++ 4.0.2 using -O3 and run on a Linux box with Intel Xeon 2.4 GHz. I can give more info on this if anyone needs more info. comparAccuracy calls several statistical distribution functions over a range of values, comparing results with what similar routines from the R-project give. The output gives the maximum abs difference and maximum of relative absolute difference between boost::math and R. The results were: norm pdf max diff: 5.55112e-17 norm pdf max rel diff: 2.20594e-16 t pdf max diff: 1.11022e-16 t pdf max rel diff: 1.17145e-15 norm cdf max diff: 1.11022e-16 norm cdf max rel diff: 2.55269e-15 t cdf max diff: 1.59317e-14 t cdf max rel diff: 3.18634e-14 F pdf max diff: 1.77636e-15 F pdf max rel diff: 7.05565e-14 F cdf max diff: 5.55112e-16 F cdf max rel diff: 1.56512e-14 chi2 pdf max diff: 5.55112e-17 chi2 pdf max rel diff: 4.26471e-14 chi2 cdf max diff: 3.46945e-16 chi2 cdf max rel diff: 6.27547e-14 exp pdf max diff: 0 exp pdf max rel diff: 0 exp cdf max diff: 0 exp cdf max rel diff: 0 gamma pdf max diff: 2.77556e-17 gamma pdf max rel diff: 3.84987e-15 gamma cdf max diff: 1.11022e-16 gamma cdf max rel diff: 1.85721e-15 lognorm pdf max diff: 1.11022e-16 lognorm pdf max rel diff: 2.21914e-16 lognorm cdf max diff: 1.11022e-16 lognorm cdf max rel diff: 1.43305e-15 cauchy pdf max diff: 0 cauchy pdf max rel diff: 0 cauchy cdf max diff: 1.11022e-16 cauchy cdf max rel diff: 2.2064e-16 weibull pdf max diff: 1.11022e-16 weibull pdf max rel diff: 2.22042e-16 weibull cdf max diff: 0 weibull cdf max rel diff: 0 norm quantile max diff: 8.88178e-16 norm quantile max rel diff: 4.00185e-16 t quantile max diff: 4.07907e-14 t quantile max rel diff: inf F quantile max diff: 1.11022e-15 F quantile max rel diff: 1.08231e-15 F quantile max diff: 1.11022e-15 F quantile max rel diff: 1.08231e-15 chi2 quantile max diff: 4.26326e-14 chi2 quantile max rel diff: 3.92646e-16 exp quantile max diff: 5.55112e-17 exp quantile max rel diff: 2.11093e-16 gamma quantile max diff: 3.55271e-15 gamma quantile max rel diff: 3.56265e-16 lognorm quantile max diff: 1.42109e-14 lognorm quantile max rel diff: 9.10487e-16 cauchy quantile max diff: 2.17426e-12 cauchy quantile max rel diff: 0.104537 weibull quantile max diff: 1.11022e-16 weibull quantile max rel diff: 2.0992e-16 uniform pdf max diff: 0 uniform pdf max rel diff: 0 uniform cdf max diff: 0 uniform cdf max rel diff: 0 uniform quantile max diff: 0 uniform quantile max rel diff: 0 beta pdf max diff: 1.11022e-15 beta pdf max rel diff: 3.14562e-15 beta cdf max diff: 4.44089e-16 beta cdf max rel diff: 2.9679e-15 beta quantile max diff: 2.22045e-16 beta quantile max rel diff: 8.38281e-16 binom pdf max diff: 2.08167e-17 binom pdf max rel diff: 0.00011212 binom cdf max diff: 9.4369e-16 binom cdf max rel diff: 3.46145e-14 poisson pdf max diff: 1.04083e-17 poisson pdf max rel diff: 1.05393e-13 poisson cdf max diff: 0 poisson cdf max rel diff: 0 nbinom pdf max diff: 1.38778e-17 nbinom pdf max rel diff: 1.16396e-13 nbinom cdf max diff: 8.88178e-16 nbinom cdf max rel diff: 2.37872e-15 lgamma max diff: 1.13687e-13 lgamma max rel diff: 0.211576 tgamma max diff: 7.43092e+142 tgamma max rel diff: 9.15759e-14 A few results need comment. ->t quantile max diff: 4.07907e-14 ->t quantile max rel diff: inf Here division by 0 gives inf. max diff is quite acceptable though. Some guards should be placed to handle 0 or near 0 but I haven't done it. I'm not worried about the accuracy but anyone may feel free to remedy this. ->cauchy quantile max diff: 2.17426e-12 ->cauchy quantile max rel diff: 0.104537 Here we are almost surely again dividing by something near 0. Similar comments as above apply. ->lgamma max diff: 1.13687e-13 ->lgamma max rel diff: 0.211576 Same old story ->tgamma max diff: 7.43092e+142 ->tgamma max rel diff: 9.15759e-14 At the higher end of the range tgamma gives such high numbers, that max diff doesn't tell us much. max rel diff gives acceptable results though. No quantile comparisons are given for the discrete distributions since R always returns integers but boost::math gives fractional values. It depends on application what you want obviously but giving results within the range (discrete in this case) is in some sense more natural. All in all I am quite happy with the results. Obviously R could be wrong but it's use is widespread in the Statistics community so it should be very well tested. comparSpeed calls a few statistical distribution functions repeatedly (see code), timing the loop, repeating the same thing for the R version. The output is the time in second: N repeats = 5000000 R norm pdf: 1.31 boost::math norm pdf: 1.1 R norm cdf: 1.49 boost::math norm cdf: 0.27 R norm quantile: 0.75 boost::math norm quantile: 1.83 R lgamma: 4.02 boost::math lgamma: 7.4 N repeats = 1000000 R t pdf: 0.91 boost::math t pdf: 4.85 R t cdf: 0.17 boost::math t cdf: 0.03 R t quantile: 12.88 boost::math t quantile: 56.15 R F pdf: 0.06 boost::math F pdf: 0.05 R F cdf: 0.07 boost::math F cdf: 0.85 R F quantile: 18.59 boost::math F quantile: 54.54 The results vary somewhat but this would be acceptable for my applications. quantile functions seem to be somewhat slower in boost::math than in R. Maybe they're more accurate. More detailed analysis would be called for if people are concerned about this. The 2 test programs are attached to this message. If they're somehow scrambled, anyone can write to me and ask for them. Last but certainly not the least, my compliments to the authors of boost::math ;-) cheers Arnaldur Gylfason, deCODE genetics

2 1

Regex match problem
by Kalin Nakov 25 Apr '07

25 Apr '07

Hi, I experience big troubles with regex matching. The code below should obviously match 'abc', but does not, because it tries to match the whole string. If r='abcdefg', then match is successfull, but if it is only 'abc' and input is 'abcdef', the match always fails as described below. boost::match_results<std::string::const_iterator> m; std::string inp = "abcdefg"; boost::regex r = "abc"; if (!boost::regex_match(inp, m, r, boost::match_default)) // always fails, don't know why How can I successfully match 'abc' agains input 'abcdefg'? Thank you very much in advance, Kalin

2 1

Re: [Boost-users] Regex match problem
by KLINIK Markus SDI2-G (AREVA NP GmbH) 25 Apr '07

25 Apr '07

You are right, regex_match tries to match the whole input string. What you want is regex_search. Best Regards Markus

1 0

Serialization and filesystem::path
by Hughes, James 25 Apr '07

25 Apr '07

Hello all, Am I right in thinking that I need to provide serialization support for filesystem::path myself because it is not built in to the serialization libraries? If that is the case, what is the best way to go about it? I had thought it was built in originally, but not been able to get things to compile - have I missed something? James This message (including any attachments) contains confidential and/or proprietary information intended only for the addressee. Any unauthorized disclosure, copying, distribution or reliance on the contents of this information is strictly prohibited and may constitute a violation of law. If you are not the intended recipient, please notify the sender immediately by responding to this e-mail, and delete the message from your system. If you have any questions about this e-mail please notify the sender immediately.

2 1

[std][msvc++][vs2005sp1] unable to open iostream with greek letters in path name
by TOL Tsagarakis Manos 25 Apr '07

25 Apr '07

Hello everyone, I know this is not a boost related issue, but I hope someone has a solution. The question is very simple and should not even be asked: “How can I open a file using std::iostreams, if the files full path, contains Greek letters (or probably any other language except English)?” The test case below shows part of the problem as it successfully opens the file using std::oftream and writes some Greek in it. Then I can read back its contents using std::ifstream. Now the file name in windows explorer or command line should be "ελληνικό όνομα αρχείου.txt" (in case you can't read this, it is "greek file name.txt"). But instead it becomes "åëëçíéêü üíïìá áñ÷åßïõ.txt" Actually when this is done by a full MFC application it is impossible to open the file! I get assertion failure! I use MS Visual Studio 2005 sp1 and std lib and CRT that comes with it. Has anyone faced this problem before? Do I have to do something special in order to be able to use standard iostreams library with Greek file names? Thank you in advanced, Emmanuil Tsagarakis // test_greek_filenames.cpp compile with: /EHsc #include <iostream> #include <fstream> #include <iterator> #include <algorithm> #include <string> #include <vector> #include <assert.h> using namespace std; namespace { struct string_by_line : std::string { }; std::istream & operator >> (std::istream & is, string_by_line & strLine) { getline(is, strLine); return is; } const string s_strFileName("ελληνικό όνομα αρχείου.txt"); } int main(int argc, char **argv) { ofstream ofs; ofs.open(s_strFileName.c_str()); assert(ofs.good()); ofs << "Writing to file ..." << endl; ofs << "Ελληνικά περιεχόμενα στο αρχείο (greek file contents) ..." << endl; ofs.close(); ifstream ifs; ifs.open(s_strFileName.c_str()); assert(ifs.good()); std::vector<std::string> vecContents; std::copy(std::istream_iterator<string_by_line>(ifs ), std::istream_iterator<string_by_line>(), std::back_inserter(vecContents)); for (size_t i=0; i<vecContents.size(); ++i) cout << vecContents.at(i) << endl; ifs.close(); }

2 2