March 2007 - Boost-users - lists.boost.org

[Test] BOOST_AUTO_TEST_CASE_EXPECTED_FAILURES => exceptions?
by Sohail Somani 21 Mar '07

21 Mar '07

Hi, Does Boost.Test consider an exception as a failure? How can I make it expect an exception? Or is there another way to do this? Thanks, Sohail

2 2

iostreams crash
by Aljaz 21 Mar '07

21 Mar '07

Hello I've been working with boost iostreams filters. Since when I changed fin.push(file_source("outest.fil", ios::in | ios::binary)); to: file_source fss("outest.fil", ios::in | ios::binary); fin.push(boost::ref(fss)); file crashes when the task is done. Why is that happening? What am I doing wrong? I think it has something to do with memory allocation.. Another thing im unsure about: Should I open file each time I want to read from it, or can I open it at program start, and leave it open (program will be running for a long period) ? I hope someone responds this time Thanks for help Best regards #include <boost/iostreams/device/file.hpp> #include <boost/iostreams/filtering_stream.hpp> #include <boost/iostreams/filter/line.hpp> #include <boost/ref.hpp> #include <exception> #include <fstream> #include <ostream> #include <iostream> #include <string> #include <vector> using namespace std; using namespace boost::iostreams; namespace io = boost::iostreams; class long_line_counter : public boost::iostreams::line_filter { public: explicit long_line_counter(int max_length = 80) : max_(max_length), count_(0) { } int count() const { return count_; } vector<string> mail_list; private: std::string do_filter(const std::string& line) { if (line.size() < max_) mail_list.push_back(line); ++count_; return line; } int max_; int count_; }; int main() { try { filtering_istream fin; long_line_counter llc(80); file_source fss("outest.fil", ios::in | ios::binary); fin.push(boost::ref(llc)); fin.push(boost::ref(fss)); char buf[4012]; while ( fin.read(buf, 4012) ) { } std::cout << llc.count() << "\n"; std::cout << llc.mail_list.size() << "\n"; } catch (exception &e) { cout << "exception: " << e.what() << std::endl; } return 0; }

1 0

[BGL] Problems implementing a special DFS algorithm
by Ignaz Rutter 21 Mar '07

21 Mar '07

Hi, I'm currently trying to implement an algorithm that basically is a DFS. The description looks roughly like this: Start a DFS, when going down an edge (e.g. put it into the tree): do something when backtracking an edge (e.g. move upwards in the search tree): do something else I tried to implement this using the DFS Visitor concept. However it seems like the appropriate events are not defined in this concept. Am I missing something here? Is it possible to reuse depth_first_search here? Or should I just I go ahead and implement my own DFS? Thanks, Ignaz

1 1

tutorial or book on boost
by Vaclav 21 Mar '07

21 Mar '07

Hello, Can you please recommend a good tutorial or book on boost library for beginners? Thanks Vaclav

4 5

Uninitialised value issue with serialization of std::vector<double>
by Jonathan Cooper 21 Mar '07

21 Mar '07

Hi. In using valgrind to run memory tests of our software, we've been getting 'Conditional jump or move depends on uninitialised value(s)' errors which appear to be caused by Boost's serialization library. Specifically, they occur on loading a std::vector<double> from a text archive. A simple test program which demonstrates this behaviour is shown below, together with a snippet from valgrind's output. Incidentally, deserializing std::vector<bool> does not have this problem, and specializing boost::serialization::{save,load} for vector as for bool makes the problem go away. Any ideas? Jonathan. Boost version: 1.33.1 Valgrind version: 3.2.1 System: Fedora Core 6 g++ (GCC) 4.1.1 20061011 (Red Hat 4.1.1-30) Test program, which compiles and runs as expected: #include <boost/archive/text_oarchive.hpp> #include <boost/archive/text_iarchive.hpp> #include <boost/serialization/vector.hpp> #include <fstream> #include <cmath> #include <cassert> int main(void) { std::string archive_filename("simple_vars.arch"); // Create an ouput archive { std::ofstream ofs(archive_filename.c_str()); boost::archive::text_oarchive output_arch(ofs); std::vector<double> doubles(3); doubles[0] = 1.1; doubles[1] = 1.2; doubles[2] = 1.3; // cast to const and write output_arch << static_cast<const std::vector<double>&>(doubles); } { // Create an input archive std::ifstream ifs(archive_filename.c_str(), std::ios::binary); boost::archive::text_iarchive input_arch(ifs); std::vector<double> doubles; // read the archive input_arch >> doubles; // test values assert(doubles.size() == 3u); assert(fabs(doubles[0] - 1.1) < 1e-12); assert(fabs(doubles[1] - 1.2) < 1e-12); assert(fabs(doubles[2] - 1.3) < 1e-12); } return 0; } Excerpt from valgrind stack trace: ==10018== Conditional jump or move depends on uninitialised value(s) ==10018== at 0x403C0ED: boost::archive::detail::basic_iarchive::reset_object_address(void const*, void const*) (in /usr/lib/libboost_serialization.so.1.33.1) ==10018== by 0x804E23E: boost::serialization::stl::archive_input_seq<boost::archive::text_iarchive, std::vector<double, std::allocator<double> > >::operator()(boost::archive::text_iarchive&, std::vector<double, std::allocator<double> >&) (collections_load_imp.hpp:54) ... ==10018== by 0x80506A7: boost::archive::text_iarchive& boost::archive::detail::interface_iarchive<boost::archive::text_iarchive>::operator>><std::vector<double, std::allocator<double> > >(std::vector<double, std::allocator<double> >&) (interface_iarchive.hpp:76) ==10018== by 0x804B7F2: main (serialize_bug.cpp:40) -- Jonathan Cooper MSN: msn(a)jonc.me.uk www: jonc.me.uk/ I never learned anything while I was talking - Larry King

1 0

[timer] How to change format of progres_display
by Seweryn Habdank-Wojewódzki 21 Mar '07

21 Mar '07

Hi Is there any easy way to change format in progress_display? Change: - Length of axis; - Number of thicks; - Labels on axis e.g. "A", "B", "C" or 1.0, 3.14, 5.12. Motivation for changing labels. If I have strings in vector and for me better is not to use 20%, but to use letters. I can take the first letter(s) of word on position "20%" and display it as label. Is it possible? Regards. -- |\/\/| Seweryn Habdank-Wojewódzki \/\/

2 2

BGL graph library biconnected components
by Greg Reynolds 21 Mar '07

21 Mar '07

Hi, I have been using "biconnected_components.hpp" for quite a while now. However, I have come across some graphs for which it finds far too many components... Here is my usage: //////////////////////////// Graph g; // initialise g and give edges an index std::vector< int > vec_bcc_nums( num_edges(g) ); iterator_property_map< std::vector< int >::iterator, property_map<Graph, edge_index_t>::type > mapEdgeComponent( vec_bcc_nums.begin() ); int num = biconnected_components(g, mapEdgeComponent); /////////////////////////// I'm afraid I can't post an example of the graphs (they are really big and I signed an NDA). What I would like to know is: a) Is the code above sensible, is there something subtly wrong with it? b) Has anyone else used the biconnected components code in anger on largish (> 300 vertices) graphs? Thanks, Greg

2 3

resetting FUSION_MAX_MAP_SIZE
by Christian Henning 20 Mar '07

20 Mar '07

Hi there, I'm experiencing problems when setting FUSION_MAX_MAP_SIZE to 20. I'm getting a lot of compiler errors: #define FUSION_MAX_MAP_SIZE 20 #include <string> #include <boost/static_assert.hpp> #include <boost/fusion/sequence.hpp> using namespace std; using namespace boost; // Address contact namespace address_fields { struct _address_1_; struct _address_2_; struct _city_; struct _state_; struct _zip_; struct _country_; struct _phone_1_; struct _phone_2_; struct _email_1_; struct _email_2_; struct _fax_; struct _other_; } typedef fusion::map< fusion::pair< address_fields::_address_1_, string > , fusion::pair< address_fields::_address_2_, string > , fusion::pair< address_fields::_city_ , string > , fusion::pair< address_fields::_state_ , string > , fusion::pair< address_fields::_zip_ , string > , fusion::pair< address_fields::_country_ , string > , fusion::pair< address_fields::_phone_1_ , string > , fusion::pair< address_fields::_phone_2_ , string > , fusion::pair< address_fields::_email_1_ , string > , fusion::pair< address_fields::_email_2_ , string > , fusion::pair< address_fields::_fax_ , string > , fusion::pair< address_fields::_other_ , string > > address_type; int _tmain(int argc, _TCHAR* argv[]) { return 0; } c:\boost\boost\fusion\sequence\container\map\map.hpp(41) : error C2977: 'boost::fusion::vector' : too many template arguments c:\boost\boost\fusion\sequence\container\vector\vector.hpp(126) : see declaration of 'boost::fusion::vector' c:\boost\boost\fusion\sequence\container\map\map.hpp(70) : see reference to class template instantiation 'boost::fusion::map<T0,T1,T2,T3,T4,T5,T6,T7,T8,T9,T10,T11,T12,T13,T14,T15,T16,T17,T18,T19>' being compiled c:\boost\boost\fusion\sequence\container\map\detail\map_lookup.hpp(31) : error C3203: 'vector' : class template invalid as template argument for template parameter 'Vector', expected a real type c:\boost\boost\fusion\sequence\container\map\detail\map_lookup.hpp(32) : see reference to class template instantiation 'boost::fusion::map<T0,T1,T2,T3,T4,T5,T6,T7,T8,T9,T10,T11,T12,T13,T14,T15,T16,T17,T18,T19>::meta_find_impl<Key,dummy>' being compiled c:\boost\boost\fusion\sequence\container\map\detail\map_lookup.hpp(37) : error C3203: 'vector' : class template invalid as template argument for template parameter 'Vector', expected a real type [snip] Any ideas, Christian

2 2

[spirit] YAML parser
by Christian Henning 20 Mar '07

20 Mar '07

Hi there, my project has come to stage where I would need to create some test data (fixtures). I would like to do it similar to Ruby where you have some yaml files containing the data. So, my question is if there is a parser using spirit for the yaml format? Thanks in advance, Christian

2 1

iostreams skip bug
by Phillip Seaver 20 Mar '07

20 Mar '07

I ran into a problem when using restrict on a file_source. The problem was that it decided that the stream wasn't seekable and called the version of skip that reads a character at a time until it gets to the requested offset, instead of just calling seek. The code was similar to this: namespace io = boost::iostreams; std::string filename; io::filtering_istream rstr(io::restrict(io::file_source(filename), 10, 100)); I've attached a patch that solved my problem. It was checking to see if the device was seekable (which means input_seekable and output_seekable), but since it was an input device, it was just input_seekable. Phillip Seaver --- skip.hpp.orig 2007-03-20 16:20:55.253000000 -0400 +++ skip.hpp 2007-03-20 16:11:51.096750000 -0400 @@ -67,7 +67,17 @@ void skip(Device& dev, stream_offset off) { typedef typename mode_of<Device>::type mode; - detail::skip(dev, off, is_convertible<mode, seekable>()); + typedef mpl::or_< + mpl::and_< + is_convertible<mode, input>, + is_convertible<mode, input_seekable> + >, + mpl::and_< + is_convertible<mode, output>, + is_convertible<mode, output_seekable> + > + > can_seek; + detail::skip(dev, off, can_seek()); } template<typename Filter, typename Device>

1 0