March 2008 - Boost-users - lists.preview.boost.org

[Proto] review period extended to end March 28th
by Hartmut Kaiser 16 Mar '08

16 Mar '08

Hi all, There have been several requests to extend the Proto review period. For this reason the review now will end on March 28th. This should give you sufficient time to have a closer look at the library and to prepare a review. The original announcement follows: The review of Eric Nieblers Proto library starts today, March 1st 2008, and will end on March 14th. I really hope to see your vote and your participation in the discussions on the Boost mailing lists! --------------------------------------------------- About the library: Proto is a framework for building Domain Specific Embedded Languages in C++. It provides tools for constructing, type-checking, transforming and executing expression templates. More specifically, Proto provides: * An expression tree data structure. * Operator overloads for building the tree from an expression. * Utilities for defining the grammar to which an expression must conform. * An extensible mechanism for immediately executing an expression template. * An extensible set of tree transformations to apply to expression trees. * A mechanism for giving expressions additional behaviors and members. Documentation is here: http://boost-sandbox.sourceforge.net/libs/proto Download proto.zip from here: http://www.boost-consulting.com/vault/index.php?directory=Template%20Metapro gramming Proto is a very important infrastructure library, IHMO. It has been used as the backbone for several other library writing efforts already, such as Xpressive and the rewrite of Spirit (there it has been used for all three parts: the parser, the lexer and the generator modules), and it is a key enabling technology in the long-planned unification of Phoenix and the Lambda Library. --------------------------------------------------- Please always state in your review, whether you think the library should be accepted as a Boost library! Additionally please consider giving feedback on the following general topics: - What is your evaluation of the design? - What is your evaluation of the implementation? - What is your evaluation of the documentation? - What is your evaluation of the potential usefulness of the library? - Did you try to use the library? With what compiler? Did you have any problems? - How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? - Are you knowledgeable about the problem domain? Regards Hartmut Review Manager

3 2

compiling with an actual compiler (gcc 4.3)
by Peter Pein 16 Mar '08

16 Mar '08

Dear group, I did a search for "4.3" and did not get a result. If the following question has been asked before, please forgive me and point me to the correct thread. I had problems compiling gnash-0.8.2. using Fedora 8 and gcc 4.3. The error message has been: g++-4.3 -DHAVE_CONFIG_H -I. -I.. -DPLUGINSDIR=\"/usr/lib/gnash/plugins\" -I.. -I. -I.. -I../libamf -I../server -I../server/swf -pthread -I/usr/include/libxml2 -I/usr/include/glib-2.0 -I/usr/lib64/glib-2.0/include -pthread -I/usr/include/gstreamer-0.10 -I/usr/include/glib-2.0 -I/usr/lib64/glib-2.0/include -I/usr/include/libxml2 -I/usr/include -I../libltdl -O3 -m64 -mfpmath=sse -mtune=athlon64 -msse2 -mmmx -m3dnow -fexpensive-optimizations -fgcse -ffast-math -pthread -pthread -W -Wall -Wcast-align -Wcast-qual -Wpointer-arith -Wreturn-type -fvisibility-inlines-hidden -MT libgnashbase_la-FLVParser.lo -MD -MP -MF .deps/libgnashbase_la-FLVParser.Tpo -c FLVParser.cpp -fPIC -DPIC -o .libs/libgnashbase_la-FLVParser.o In file included from /usr/include/boost/date_time/gregorian/gregorian_io.hpp:12, from /usr/include/boost/date_time/gregorian/conversion.hpp:23, from /usr/include/boost/date_time/gregorian/gregorian.hpp:21, from /usr/include/boost/date_time/posix_time/time_formatters.hpp:12, from /usr/include/boost/date_time/posix_time/posix_time.hpp:24, from ../libamf/element.h:31, from ../libamf/amf.h:30, from FLVParser.cpp:25: /usr/include/boost/date_time/date_facet.hpp:560: error: declaration of 'void boost::date_time::date_input_facet<date_type, CharT, InItrT>::special_values_parser(boost::date_time::special_values_parser<date_type, charT>)' /usr/include/boost/date_time/special_values_parser.hpp:34: error: changes meaning of 'special_values_parser' from 'class boost::date_time::special_values_parser<date_type, charT>' make[2]: *** [libgnashbase_la-FLVParser.lo] Fehler 1 make[2]: Leaving directory `/usr/local/src/gnash-0.8.2/libbase' make[1]: *** [all-recursive] Fehler 1 make[1]: Leaving directory `/usr/local/src/gnash-0.8.2' make: *** [all] Fehler 2 (well, its partly german; "Fehler" means "error". Is this an error inside boost? Or is it possible that gnash-sources define a prototype in a wrong manner? I tried to compile boost and used F77=gfortran-4.3 CC="ccache gcc-4.3" CXX="ccache g++-4.3" CFLAGS="-O3 -mfpmath=sse -mtune=athlon64 -msse2 -mmmx -m3dnow -fexpensive-optimizations -fgcse -ffast-math -m64" CXXFLAGS="-O3 -mfpmath=sse -mtune=athlon64 -msse2 -mmmx -m3dnow -fexpensive-optimizations -fgcse -ffast-math -m64" ./configure --prefix=/usr --libdir=/usr/lib64 --with-libraries=all --exec-prefix=/usr/bin to configure it and because I got messages like gcc.compile.c++ bin.v2/libs/thread/build/gcc-4.1.2/debug/threading-multi/condition.o during make, I did a make distclean, another configure as above and typed: make F77=gfortran-4.3 CC="ccache gcc-4.3" CXX="ccache g++-4.3" CFLAGS="-O3 -mfpmath=sse -mtune=athlon64 -msse2 -mmmx -m3dnow -fexpensive-optimizations -fgcse -ffast-math -m64" CXXFLAGS="-O3 -mfpmath=sse -mtune=athlon64 -msse2 -mmmx -m3dnow -fexpensive-optimizations -fgcse -ffast-math -m64" and got the same log telling me there is something compiled for the prepredecessor of the actual compiler. Do I have to edit all the Makefiles by hand? If yes, what is the ./configure script for? Or do I have to use the old old gcc-4.1 ("old" would have been gcc-4.2) ? Greetings and TIA from a hopelessly lost Peter P.S.: my mail-client told me it could not send it because I need a valid email address (nonsense, because I tried it first with my mostly used address). If this message comes twice, I apologize.

5 4

[multi_index] GCC optimization warns about uninitialized variables
by cp now 16 Mar '08

16 Mar '08

Hello. I have a multi_index container that stores (and owns) pointers. It has 4 indexes that order themselves on member functions. I get error-free and warning-free compiles as long as optimization is turned off (-O0). I get the warning below at all levels of optimization (1 through 3). I've stripped the code down to just what seems to be generating the error. --- BEGIN test.cpp --- #include <iostream> #include <typeinfo> #include <map> #include <boost/ref.hpp> #include <boost/multi_index_container.hpp> #include <boost/multi_index/sequenced_index.hpp> #include <boost/multi_index/ordered_index.hpp> #include <boost/multi_index/key_extractors.hpp> using namespace boost::multi_index; typedef boost::reference_wrapper<const std::type_info> TypeInfo; #if 0 bool operator<(TypeInfo t1, TypeInfo t2) { const std::type_info& t = t1; return t.before(t2); } #endif namespace std { template<> struct less<TypeInfo> { bool operator()(const TypeInfo& t1, const TypeInfo& t2) { const std::type_info& t = t1; return t.before(t2); } }; } struct NodeBase { std::string getID() const {return "blah";} TypeInfo getresulttypeinfo() const {return TypeInfo(typeid(int));} std::string name() const {return "name";} }; typedef multi_index_container< const NodeBase*, indexed_by< sequenced<> , ordered_unique< const_mem_fun<NodeBase, std::string, &NodeBase::getID> > , ordered_non_unique< const_mem_fun<NodeBase, TypeInfo, &NodeBase::getresulttypeinfo> > , ordered_non_unique< const_mem_fun<NodeBase, std::string, &NodeBase::name> > > > NodeMultiIndex; int main() { NodeMultiIndex c; NodeMultiIndex::nth_index<0>::type& index = c.get<0>(); index.push_back(new NodeBase()); std::cout << *index.begin() << std::endl; return 0; } --- END test.cpp --- Compiling with G++ 4.1.1 with full optimizations and full warnings: g++-4.1.1 -I../eclipselibs/boost -O3 -Wall -c -fmessage-length=0 test.cpp ../eclipselibs/boost/boost/multi_index/ordered_index.hpp: In member function 'boost::multi_index::detail::ordered_index_node<typename SuperMeta::type::node_type>* boost::multi_index::detail::ordered_index<KeyFromValue, Compare, SuperMeta, TagList, Category>::insert_(typename boost::call_traits<typename boost::multi_index::detail::ordered_index_node<typename SuperMeta::type::node_type>::value_type>::param_type, boost::multi_index::detail::ordered_index_node<typename SuperMeta::type::node_type>*) [with KeyFromValue = boost::multi_index::const_mem_fun<NodeBase, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, &NodeBase::getID>, Compare = std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, SuperMeta = boost::multi_index::detail::nth_layer<2, const NodeBase*, boost::multi_index::indexed_by<boost::multi_index::sequenced<boost::multi_index::tag<mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na> >, boost::multi_index::ordered_unique<boost::multi_index::const_mem_fun<NodeBase, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, &NodeBase::getID>, mpl_::na, mpl_::na>, boost::multi_index::ordered_non_unique<boost::multi_index::const_mem_fun<NodeBase, boost::reference_wrapper<const std::type_info>, &NodeBase::getresulttypeinfo>, mpl_::na, mpl_::na>, boost::multi_index::ordered_non_unique<boost::multi_index::const_mem_fun<NodeBase, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, &NodeBase::name>, mpl_::na, mpl_::na>, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na, mpl_::na>, std::allocator<const NodeBase*> >, TagList = boost::mpl::vector0<mpl_::na>, Category = boost::multi_index::detail::ordered_unique_tag]': ../eclipselibs/boost/boost/multi_index/ordered_index.hpp:551: warning: 'inf$side' may be used uninitialized in this function I can eliminate this warning (for all levels of optimization) with the patch below. Index: boost/multi_index/ordered_index.hpp =================================================================== RCS file: /cvsroot/boost/boost/boost/multi_index/ordered_index.hpp,v retrieving revision 1.12.2.3 diff -u -r1.12.2.3 ordered_index.hpp --- boost/multi_index/ordered_index.hpp 26 Feb 2007 09:16:36 -0000 1.12.2.3 +++ boost/multi_index/ordered_index.hpp 16 Mar 2008 01:40:02 -0000 @@ -549,6 +549,7 @@ node_type* insert_(value_param_type v,node_type* x) { link_info inf; + inf.side = to_left; inf.pos = 0; if(!link_point(key(v),inf,Category())){ return node_type::from_impl(inf.pos); } I imagine something similar would have to be done for the other insert functions, though my code probably doesn't use them and so the warnings aren't generated at that line. Alternatively, the link_info struct could have a default constructor that initializes the members. Or link_point() etc. could return tuples, eliminating the need for pass-by-reference, etc. Is there an easier workaround than patching the library header? Or, worse, am I doing something wrong? Thanks V

2 1

[serialization] Cannot increase version of stored objects
by Robert Bielik 15 Mar '08

15 Mar '08

I need to update the version certain objects in my structure, but I cannot get the version number incremented. Scenario: I have a class MyClass which have several entries of InternalClass objects (which is the one I want a new version of): class MyClass { ... template<class Archive> void serialize(Archive & ar, const unsigned int version) { ar & BOOST_SERIALIZATION("MyInternals", m_vector_of_internal_class_objects); } std::vector<InternalClass> m_vector_of_internal_class_objects; class InternalClass { template<class Archive> void save(Archive & ar, const unsigned int version) const { ar & BOOST_SERIALIZATION("data", m_data); } template<class Archive> void load(Archive & ar, const unsigned int version) { if (version == 0) { old_data_t tmp; ar & BOOST_SERIALIZATION("data", tmp); m_data = convertToNewData(tmp); } else { BOOST_CHECK(version == 1); ar & BOOST_SERIALIZATION("data", m_data); } } BOOST_SERIALIZATION_SPLIT_MEMBER() new_data_t m_data; } } Then in the body (.cpp) I define: BOOST_CLASS_VERSION(MyClass::InternalClass, boost::serialization::object_class_info) BOOST_CLASS_VERSION(MyClass::InternalClass, 1) But when saving to the XML archive file, I still get version="0" for the InternalClass objects (?) /R

2 3

function template overload resolutions
by Hicham Mouline 15 Mar '08

15 Mar '08

Hello As functions templates cannot be partially specialized, Sutter(05) advices that one can use overload resolution instead: I have // header N.h namespace A { namespace B { namespace C { template<typename cT, typename tT, int n> void make(const cT&, const tT&, const Head<n>& h) {...} } } } // header 3.h namespace A { namespace B { namespace C { template<typename cT, typename tT> void make(const cT&, const tT&, const Head<3>& h) {...} } } } // header 2.h namespace A { namespace B { namespace C { template<typename cT, typename tT> void make(const cT&, const tT&, const Head<2>& h) {...} } } } // Translation unit #include N.h #include 3.h #include 2.h //in this order A::B::C::make(...) The overload resolution is done on the Head<n>&, Head<2>&, Head<3>... Is it sensitive to order of inclusion? rds,

1 0

dll or not dll or what?
by Martina Janter 15 Mar '08

15 Mar '08

Hello! being a beginner in C++ and boost I set up the stuff for use with VC++ 2005 Express just today. After some problems I finally managed to get it link, but it's still not fully the way I want. There are three libs: boost...dll, libboost...lib and libboost...s...lib. My questions: -What's the difference between the second and the third one? -Why does building with the IDE link against the second and building on the console (/ link /LIBPATH:...) link against the third type? I looked into auto_link.hpp and the IDE seems to set _DLL but boost doesn't define BOOST_DYN_LINK. But I can't find out why. This leads to question 3: -How to link so the dll will be used? That questions might be a good deal compiler specific but since auto_link.hpp is involved I hope here's still the right place to ask. Thank you in advance! AJ Martin

2 1

[multi_array] Assertion failure in multi_array<...>::resize
by Rainer Deyke 15 Mar '08

15 Mar '08

The following minimal test program produces an assertion failure. Is this a bug in boost::multi_array or a bug in the test program? #include "boost/multi_array.hpp" int main() { boost::multi_array<bool, 1> a; a.resize(boost::extents [boost::multi_array_types::extent_range(1, 2)]); return 0; } -- Rainer Deyke - rainerd(a)eldwood.com

1 0

[interprocess] Where to get it?
by Boris 14 Mar '08

14 Mar '08

Is the CVS the only place to get the source code of Boost.Interprocess? I thought I would find a zip archive at http://www.boost-consulting.com/vault/ at least so I don't need to pick files one by one from the CVS. Any other place where I can download the current version of Boost.Interprocess quickly? I know the documentation at http://igaztanaga.drivehq.com/interprocess/ but there is no link either? :-/ Boris

2 4

[Mutex] behaviour??
by Yasmina Berber 14 Mar '08

14 Mar '08

Hi, everybody I have a problem to understand mutex behaviour in a server client program . basically there are two threads that execute different code in two different classes but manipulate the same resource: fisrt Class is Subscriber with these members : public: ACE_Unbounded_Set<client_reference> m_set; typedef boost::recursive_mutex Mutex; typedef boost::recursive_mutex::scoped_lock Lock; Mutex mutex_set; and methods: Subscribe (client_reference) : inserts (add) a reference client to the set { ... Lock theScope( mutex_set); m_set.insert (client_reference) ... } Unsubscribe(client_reference) : removes a reference client from the set { ... Lock theScope( mutex_set); m_set.remove (client_reference) } the second class is Manager that has a shared pointer to Subscriber (SubscriberPtr) and does a specific opeartion on each client_reference inside a for loop in the method Manage() { boost::recursive_mutex::scoped_lock scoped_lock(subscriberPtr->mutex_set) for (Subscriber::Set_Iterator iter (subscriberPtr->m_set); iter.next (...) != 0 ; iter.advance ()) { specific opeartion on each client reference } } What happens is that a client can be removed (call to Unsubscribe()) by the first thread during the second one is executing manage() : the set is not well protected and this causes of course seg fault. I probably do something wrong, If you need more details please let me know. Regards.

2 1

Boost.Regex and static ICU libraries
by a86yes 14 Mar '08

14 Mar '08

Hello, The ICU4C readme says that the define U_STATIC_IMPLEMENTATION should be defined in user code that links against ICU's static libraries. Grepping the entire Boost, I do not see it. Is linking against ICU static libraries supported at all and if yes, where is the best place to define this macro when building Boost.Regex? _________________________________________________________________ Сдобий се със следващата генерация Windows Live услуги сега! http://get.live.com

2 3