- Boost - lists.boost.org

[Boost-bugs] [ boost-Bugs-1166401 ] remove_edge fails with gcc 3.4.3, likely due to libstdc++
by SourceForge.net 19 Mar '05

19 Mar '05

Bugs item #1166401, was opened at 2005-03-18 21:50 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=1166401&group_… Category: graph Group: None Status: Open Resolution: None Priority: 5 Submitted By: Nobody/Anonymous (nobody) Assigned to: Jeremy Siek (jsiek) Summary: remove_edge fails with gcc 3.4.3, likely due to libstdc++ Initial Comment: In certain situations a call to remove_edge fails to accomplish the removal. I believe this is dues to a problem with std::equal_range in libstdc++ in gcc 3.4.3 but until this is fixed a workaround in boost that does not use equal_range could fix this problem. The attached test program shows the problem. On line 68 removing p24 is ok but removing p12 silently fails, that is after completion the output shows 3 vertices and 3 edges remaining in the graph while it should have only 3 vertices and 2 edges left. The point of failure is the following call stack: (gdb) where #0 std::equal_range<std::_List_iterator<boost::detail::sei_<void*, std::_List_iterator<boost::list_edge<void*, boost::property<BoostEdgeContentType, int*, boost::no_property> > >, boost::property<BoostEdgeContentType, int*, boost::no_property> > >, boost::detail::sei_<void*, std::_List_iterator<boost::list_edge<void*, boost::property<BoostEdgeContentType, int*, boost::no_property> > >, boost::property<BoostEdgeContentType, int*, boost::no_property> > > (__first= {_M_node = 0x804f098}, __last={_M_node = 0x804f098}, __val=@0xbfffecc0) at /apps/utkej_projects/Argonne/gcc-3.4.3_inst/lib/gcc/i686-pc-linux-gnu/3.4.3/../../../../include/c++/3.4.3/bits/stl_algo.h:3814 #1 0x0804a8e9 in boost::edge_range<boost::detail::adj_list_gen<boost::adjacency_list<boost::listS, boost::listS, boost::bidirectionalS, boost::property<BoostVertexContentType, int*, boost::no_property>, boost::property<BoostEdgeContentType, int*, boost::no_property>, boost::no_property, boost::listS>, boost::listS, boost::listS, boost::bidirectionalS, boost::property<BoostVertexContentType, int*, boost::no_property>, boost::property<BoostEdgeContentType, int*, boost::no_property>, boost::no_property, boost::listS>::config, boost::bidirectional_graph_helper_with_property<boost::detail::adj_list_gen<boost::adjacency_list<boost::listS, boost::listS, boost::bidirectionalS, boost::property<BoostVertexContentType, int*, boost::no_property>, boost::property<BoostEdgeContentType, int*, boost::no_property>, boost::no_property, boost::listS>, boost::listS, boost::listS, boost::bidirectionalS, boost::property<BoostVertexContentType, int*, boost::no_property>, boost::property<BoostEdgeContentType, int*, boost::no_property>, boost::no_property, boost::listS>::config> > (u=0x804f098, v=0x804f068, g_=@0xbfffeea0) at /apps/utkej_projects/Argonne/boost_1_32_0/boost/graph/detail/adjacency_list.hpp:1485 #2 0x08049d8e in boost::bidirectional_graph_helper_with_property<boost::detail::adj_list_gen<boost::adjacency_list<boost::listS, boost::listS, boost::bidirectionalS, boost::property<BoostVertexContentType, int*, boost::no_property>, boost::property<BoostEdgeContentType, int*, boost::no_property>, boost::no_property, boost::listS>, boost::listS, boost::listS, boost::bidirectionalS, boost::property<BoostVertexContentType, int*, boost::no_property>, boost::property<BoostEdgeContentType, int*, boost::no_property>, boost::no_property, boost::listS>::config>::remove_edge (this=0xbfffeea0, e= {<boost::detail::edge_base<boost::bidirectional_tag,void*>> = {m_source = 0x804f098, m_target = 0x804f068}, m_eproperty = 0x804f0d8}) at /apps/utkej_projects/Argonne/boost_1_32_0/boost/graph/detail/adjacency_list.hpp:1067 #3 0x080495aa in boost::remove_edge<boost::detail::edge_desc_impl<boost::bidirectional_tag, void*>, boost::detail::adj_list_gen<boost::adjacency_list<boost::listS, boost::listS, boost::bidirectionalS, boost::property<BoostVertexContentType, int*, boost::no_property>, boost::property<BoostEdgeContentType, int*, boost::no_property>, boost::no_property, boost::listS>, boost::listS, boost::listS, boost::bidirectionalS, boost::property<BoostVertexContentType, int*, boost::no_property>, boost::property<BoostEdgeContentType, int*, boost::no_property>, boost::no_property, boost::listS>::config> (e= {<boost::detail::edge_base<boost::bidirectional_tag,void*>> = {m_source = 0x804f098, m_target = 0x804f068}, m_eproperty = 0x804f0d8}, g_=@0xbfffeea0) at /apps/utkej_projects/Argonne/boost_1_32_0/boost/graph/detail/adjacency_list.hpp:1110 #4 0x0804906a in main () at gv2.cpp:70 (gdb) Basically the problem is in stl_algo,h when __len on 3801 is initally 2 and __val in our case equals the second element and the condition on 3810 holds true. Note that for this to happen the vertices and edges in the graph need to be created in a certain order. In any case, under the above conditions we compute at 3814 __len to be 0 and __first becomes equal to __last. I.e. we don't ever look at the second element in the list which is the one we SHOULD find. 3784 template<typename _ForwardIterator, typename _Tp> 3785 pair<_ForwardIterator, _ForwardIterator> 3786 equal_range(_ForwardIterator __first, _ForwardIterator __last, 3787 const _Tp& __val) 3788 { 3789 typedef typename iterator_traits<_ForwardIterator>::value_type 3790 _ValueType; 3791 typedef typename iterator_traits<_ForwardIterator>::difference_type 3792 _DistanceType; 3793 3794 // concept requirements 3795 // See comments on lower_bound. 3796 __glibcxx_function_requires(_ForwardIteratorConcept<_ForwardIterator>) 3797 __glibcxx_function_requires(_SameTypeConcept<_Tp, _ValueType>) 3798 __glibcxx_function_requires(_LessThanComparableConcept<_Tp>) 3799 __glibcxx_requires_partitioned(__first, __last, __val); 3800 3801 _DistanceType __len = std::distance(__first, __last); 3802 _DistanceType __half; 3803 _ForwardIterator __middle, __left, __right; 3804 3805 while (__len > 0) 3806 { 3807 __half = __len >> 1; 3808 __middle = __first; 3809 std::advance(__middle, __half); 3810 if (*__middle < __val) 3811 { 3812 __first = __middle; 3813 ++__first; 3814 __len = __len - __half - 1; 3815 } 3816 else if (__val < *__middle) Subsequently the rng returns the end/end and the edge in question never gets deleted because of edge_range(source(e, g), target(e, g), g); rng.first = std::find(rng.first, rng.second, e); if (rng.first != rng.second) remove_edge(rng.first); } in graph/detail/adjacency_list.hpp unfortunately without warning. Temporarily not using equal_range should fix it. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1166401&group_… ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Boost-bugs mailing list Boost-bugs(a)lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/boost-bugs

1 0

Anyone interested in the following boost "utility" library
by Tom Brinkman 19 Mar '05

19 Mar '05

How do you specify a string or a double, or an object as a template parameter? c++ templates are a compile time construct, so the answer is that you can't, or at least not directly. However, with the use of existing boost libraries and a little compiler magic, you can actually specify a string, double or any object as a template parameter. The following code is the proof of this concept. I would like to gather feedback from the group and submit this as a boost utility library if there is sufficient interest. //TODO //1) compile the following code as main.cpp. //2) run the application with three command line parameters. #include <vector> #include <string> #include <iostream> #include <boost/variant.hpp> std::vector<boost::variant<int, long,double,std::string> > resources; template <bool> struct resource{}; template <> struct resource<false> { template <typename INDEX, typename R> void operator()(INDEX index, R& out) { out = oost::get<R> (resources[index()]); } }; template <> struct resource<true> { template <typename T, typename R> void operator()(T in, R& out) { out = in; } }; template <typename T, typename R> inline void get(R& out) { resource<boost::is_convertible< T,R>::value> a; a(T(),out); } template <int T=-1> struct resource_wrapper { static const int value = T; int operator()(){return T;} }; template <int T=0> struct integer { static const int value = T; operator int(){return T;} }; typedef resource_wrapper<1> str1; typedef resource_wrapper<2> str2; typedef resource_wrapper<3> str3; typedef resource_wrapper<4> columns; template <typename STRING> struct funct { void operator()() { std::string str; get<STRING>(str); std::cout << str << std::endl; } }; template <typename DOUBLE> struct funct2 { void operator()() { double dbl; get<DOUBLE>(dbl); std::cout << dbl << std::endl; } }; int main(int argc, char* argv[]) { resources.resize(100); resources[str1::value] = std::string(argv[1]); resources[str2::value] = std::string(argv[2]); resources[str3::value] = std::string(argv[3]); resources[columns::value] = 6.33; funct<str1>()(); funct<str2>()(); funct<str3>()(); funct2<columns>()(); funct2<integer<10> >()(); return 0; } __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

1 0

google going open source
by Bronek Kozicki 19 Mar '05

19 Mar '05

under BSD 2.0 license http://code.google.com/ B.

8 14

test_tools.hpp problem with borland
by Robert Ramey 18 Mar '05

18 Mar '05

When compiling the following small test program with either of the borland compilers: #include <boost/test/test_tools.hpp> class base { }; int main(int argc, char *argvp[]){ BOOST_WARN_MESSAGE(true, "test warning"); } I get the following syntax error: Borland C++ 5.6 for Win32 Copyright (c) 1993, 2002 Borland c:\BoostMainCVS\libs\serialization\test\test_zmisc.cpp: Warning W8057 c:\BoostMainCVS\libs\serialization\test\test_zmisc.cpp 13: Parameter 'argc' is never used in function main(int,char * *) Warning W8057 c:\BoostMainCVS\libs\serialization\test\test_zmisc.cpp 13: Parameter 'argvp' is never used in function main(int,char * *) Error E2312 C:\BoostMainCVS\boost/test/predicate_result.hpp 60: 'base' is not an unambiguous base class of 'predicate_result::readonly_property60' in function predicate_result::readonly_property60::readonly_property60(bool) *** 1 errors in Compile *** It would seem to me that the test library needs a borland specific tweak to address this. Robert Ramey

2 1

[Boost-bugs] [ boost-Bugs-1165994 ] Annoyed warning during bjam building by Sun C
by SourceForge.net 18 Mar '05

18 Mar '05

Bugs item #1165994, was opened at 2005-03-18 18:13 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=1165994&group_… Category: None Group: None Status: Open Resolution: None Priority: 5 Submitted By: simon (atanasyan) Assigned to: Nobody/Anonymous (nobody) Summary: Annoyed warning during bjam building by Sun C Initial Comment: == Bug description: Boost version: 1.32.0 I builde bjam by boost/tools/build/jam_src/build.sh on Solaris 9 by Sun C compiler and get annoyed and potential dangerous warnings. % uname -srp SunOS 5.9 sparc % cd boost/tools/build/jam_src % export BOOST_JAM_TOOLSET_ROOT=$HOME/ws//vulcan/dest/sparc-S2/ % ./build.sh sunpro ... skipped ... .cc. bin.solaris/yyacc cc: Warning: -xarch=native has been explicitly specified, or implicitly specified by a macro option, -xarch=native on this architecture implies -xarch=v8plusb which generates code that does not run on pre UltraSPARC III processors == Proposed fix: Set the -xarch=generic. I attached the patch to this report. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1165994&group_… ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Boost-bugs mailing list Boost-bugs(a)lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/boost-bugs

1 0

[Boost-bugs] [ boost-Bugs-1165980 ] build.sh doesn't build bjam under Solaris by Sun C++
by SourceForge.net 18 Mar '05

18 Mar '05

Bugs item #1165980, was opened at 2005-03-18 17:54 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=1165980&group_… Category: None Group: None Status: Open Resolution: None Priority: 5 Submitted By: simon (atanasyan) Assigned to: Nobody/Anonymous (nobody) Summary: build.sh doesn't build bjam under Solaris by Sun C++ Initial Comment: == Bug description: Boost version: 1.32.0 I work under Solaris 9 and my SunPro compiler is in the unusual place. It looks like boost/tools/build/jam_src/build.sh can't build bjam. % uname -srp SunOS 5.9 sparc % cd boost/tools/build/jam_src % export BOOST_JAM_TOOLSET_ROOT=$HOME/ws//vulcan/dest/sparc-S2/ % /bin/sh -x build.sh BOOST_JAM_TOOLSET= BOOST_JAM_TOOLSET=sunpro + shift BOOST_JAM_OPT_JAM=-o bootstrap/jam0 BOOST_JAM_OPT_MKJAMBASE=-o bootstrap/mkjambase0 BOOST_JAM_OPT_YYACC=-o bootstrap/yyacc0 + test -r /opt/SUNWspro/bin/cc + test -r /cc <<<<<< Error BOOST_JAM_CC=cc == Proposed fix: I attach the patch to this report. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1165980&group_… ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Boost-bugs mailing list Boost-bugs(a)lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/boost-bugs

1 0

[iostreams] Reading a part of file
by Vladimir Prus 18 Mar '05

18 Mar '05

Hello, I wonder if the iostreams library allows me to get an istream that will read a part of a file. That is: I specify a file, start offset and end offset. The resulting istream begins reading at start offset and returns EOF when it arrives at end offset. I haven't found anything like this in the docs and I also don't know if such functionality makes sense for the library. I just need to read individual members of Unix static archives (.a files) and wanted to know if the library can help here. If not, I'll just read the data into strstream. Also: the POSIX mmap call allows one to specify start and end offset. However, the mapped_file_source does not allow to specify that, and documentation does not clearly explain why one would want to use that class at all. TIA, Volodya

2 3

Re: [boost] Re: [ boost-Bugs-1164057 ]boost::filesystem::existshas bugs with UNC paths
by Iain Hanson 18 Mar '05

18 Mar '05

On Thu, 2005-03-17 at 13:51 -0500, Robert Mathews wrote: > "Martin Bonner" <martin.bonner(a)pitechnology.com> wrote in message > news:E562FCEE3A42D61192880002A5FB433302D7A985@kite.pigroup.co.uk... > So, either, we've got to consider "//server" a valid path, or change the > iterator so that it only returns valid paths. I tend to think that fixing > the iterator would be a better choice. > To me, logically, //server is a valid and well formed path. It just so happens that I can't access it ( permission denied ). /ikh

1 0

[Boost-bugs] [ boost-Bugs-1165844 ] С ompilation test/minmax_element_test.cpp failed on Sunpro
by SourceForge.net 18 Mar '05

18 Mar '05

Bugs item #1165844, was opened at 2005-03-18 13:12 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=1165844&group_… Category: test Group: None Status: Open Resolution: None Priority: 5 Submitted By: simon (atanasyan) Assigned to: Beman Dawes (beman_dawes) Summary: Сompilation test/minmax_element_test.cpp failed on Sunpro Initial Comment: = Bug description: Boost version: 1.32.0 Compiler: Sun Studio 10 release (C++ 5.7) with patch 117830-01 % cd ~/boost/libs/algorithm/minmax/test % CC -library=stlport4 -c -library=stlport4 -I/home/sa154628/boost minmax_element_test.cpp minmax_element_test.cpp, line 230: Error: The function "atoi" must have a prototype. 1 Error(s) detected. == Proposed fix Only std c++ headers are included in the minmax_element_test.cpp. Therefore atoi is in the std namespace and isn't exported to global one. So we should change call atoi to std::atoi. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=107586&aid=1165844&group_… ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ Boost-bugs mailing list Boost-bugs(a)lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/boost-bugs

1 0

[ptr_container] ptr_map throws in []
by Felipe Magno de Almeida 18 Mar '05

18 Mar '05

Hi. I dont see why ptr_map throws in operator[], I think it should have the same behavior of std::map and return a default constructed object in case it doesnt find the key. Anyway, it is just a opinion. Im using ptr_map in production code and Im obligated to use find to see if the key is already in the map, if it isnt. create one. -- Felipe Magno de Almeida UIN: 2113442 email: felipe.almeida at ic unicamp br, felipe.m.almeida at gmail com, felipe at synergy com I am a C, modern C++, MFC, ODBC, Windows Services, MAPI developer from synergy, and Computer Science student from State University of Campinas(UNICAMP). To know more about: Unicamp: http://www.ic.unicamp.br Synergy: http://www.synergy.com.br current work: http://www.mintercept.com

2 4