September 2007 - Boost-users - lists.preview.boost.org

[multi-index]/shared_ptr template issue - compiles on MSVC8 but not gcc
by Mark Schlegel 20 Sep '07

20 Sep '07

Hi, I'm using boost versions 1.33.1 and 1.34.1 with gcc 4.1.2 (Linux) and boost 1.33.1 with gcc 3.3.2 on solaris and MSVC8 I'm trying to use multi_index as a container indexed by two integers in a class called MapEntry. The two indices represent a unique key number "key" and a possibly non-unique timestamp "age". To make it generic I have the data in a shared_ptr<T> in the MapEntry class because I wanted to use the aged map functionality with many classes in the shared_ptr: // the tags struct keytag { }; struct agetag { }; template <typename T> struct MapEntry { .... stripped out all constructors and methods ... int key; int age; private: boost::shared_ptr<T> payload; }; and the multi_index_container setup as a wrapper class since C++ doesn't have typedef templates because I figured I needed that to get the "T" into the MapEntry<T> .... template<typename T> class AgedMap { typedef T map_type; public: typedef multi_index_container< MapEntry<T>, indexed_by< ordered_unique< tag<keytag>, BOOST_MULTI_INDEX_MEMBER(MapEntry<T>, int, key) >, ordered_non_unique< tag<agetag>, BOOST_MULTI_INDEX_MEMBER(MapEntry<T>, int, age) > > > AGEDMAPTYPE; }; and in main() I declare the container as: AgedMap<AH_ENTRY>::AGEDMAPTYPE am; and do insert, finds and erases into it. My problem is that it seems that if I use the T template to call my flush_by_age template function : template <typename T, typename MIContainer> void flush_by_age(MIContainer& mic, int timeout) { ..... which is called from main() as (note AH_ENTRY is a very simple struct with just a single int in it and cctor/ctor/dtor for testing) flush_by_age<AH_ENTRY, AHS_MAP>(am, AGELIMIT); and use that T to define what AGE_INDEX is inside it: // this hard coded one compiles and runs on all my platforms // but loses the whole point of genericness //typedef AgedMap<AH_ENTRY>::AGEDMAPTYPE::index<agetag>::type AGE_INDEX; // this AGE_INDEX does not on gcc but DOES compile on MSVC8 typedef AgedMap<T>::AGEDMAPTYPE::index<agetag>::type AGE_INDEX; I get a compiler error at the AGE_INDEX typedef in gcc 3.3.2 (solaris 10): aged_map.cc: In function `void flush_by_age(MIContainer&, int)': aged_map.cc:159: warning: `boost::multi_index::multi_index_container<Value1, IndexSpecifierList1, Allocator1>::index<agetag>::type' is implicitly a typename and a more verbose one on gcc 4.1.2 (Fedora 6 Linux): aged_map.cc: In function ÃÂ¢void flush_by_age(MIContainer&, int)ÃÂ¢: aged_map.cc:159: error: non-template 'index' used as template aged_map.cc:159: note: use ÃÂ¢boost::multi_index::multi_index_container<MapEntry<T>, boost::multi_index::indexed_by<boost::multi_index::ordered_unique< boost::multi_index::tag<keytag, 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::member<MapEntry<T>, int, (& MapEntry<T>::key)>, mpl_::na>, boost::multi_index::ordered_non_unique< boost::multi_index::tag<agetag, 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::member<MapEntry<T>, int, (& MapEntry<T>::age)>, 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>, std::allocator<MapEntry<T> > >::template index' to indicate that it is a template aged_map.cc:159: error: too few template-parameter-lists But MSVC 8 compiles it fine and it runs. But if I hard code the class in my AGE_INDEX typedef to "<AH_ENTRY>" to replace the <T> in the flush_by_age() it compiles fine on ALL those compilers and they all run ok. I'm wanting to find out what I can possibly do because this would be a very useful class for since I currently have to put the aging information in another container where this let's me do it in one. I use unix and Windows and on Windows the compiler must be MSVC and on unix gcc. I tried making the typename explicit by adding "typename" using the hint from the solaris error but that didn't help. I'm guessing I've got some syntax typo like a missing & or const somewhere ? thanks, Mark

3 2

(no subject)
by Mark Schlegel 20 Sep '07

20 Sep '07

>> // this hard coded one compiles and runs on all my platforms >> // but loses the whole point of genericness >> //typedef AgedMap<AH_ENTRY>::AGEDMAPTYPE::index<agetag>::type AGE_INDEX; >> >> // this AGE_INDEX does not on gcc but DOES compile on MSVC8 >> typedef AgedMap<T>::AGEDMAPTYPE::index<agetag>::type AGE_INDEX; > Try adding 'typename' after 'typedef'. Yes, that was one of the things I tried before posting, it didn't help (but I didn't leave that in the posted code example) Mark

1 0

Boost unit test fails to compile with Visual Studio 8 SP1
by Sebastian Hauer 20 Sep '07

20 Sep '07

Hi, I just upgrade my visual studio 8 install to SP1 and now I am unable to compile a simple unit test. The compilation error I am getting is: cl /EHsc /W3 /Wp64 /RTCsu /Gm /Ob0 /Od /MTd /TP /DFUSE_DB_WITH_MFC /DWIN32 /D_DEBUG /D_MBCS /IY:\sys\lib\boost_1_34_1\include /Idb\src /c db\test\fuse_db-tests.cpp /Fobuild-db-vc80-mt-sgd\test\fuse_db-tests.obj /Zi /Fdbuild-db-vc80-mt-sgd\test\fuse_db-tests.obj.pdb Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86 Copyright (C) Microsoft Corporation. All rights reserved. fuse_db-tests.cpp In fuse_db_tests.cpp C:\Program Files\Microsoft Visual Studio 8\VC\INCLUDE\iterator(730) : error C2953: 'stdext::checked_array_iterator' : class template has already been defined C:\Program Files\Microsoft Visual Studio 8\VC\INCLUDE\xutility(3220) : see declaration of 'stdext::checked_array_iterator' When I compile with /showIncludes I see that the xutility header is included by: Note: including file: Y:\sys\lib\boost_1_34_1\include\boost/shared_ptr.hpp Note: including file: Y:\sys\lib\boost_1_34_1\include\boost/config.hpp Note: including file: C:\Program Files\Microsoft Visual Studio 8\VC\INCLUDE\memory Note: including file: C:\Program Files\Microsoft Visual Studio 8\VC\INCLUDE\iterator Has anyone else run into this problem? Regards, Sebastian

4 4

sp_counted_base.hpp and VxWorks
by Daniel P Furlani 20 Sep '07

20 Sep '07

The file sp_counted_base.hpp checks a bunch of macros to determine which architecture-specific include file to use. One of those checks is: #elif defined( __GNUC__ ) && ( defined( __powerpc__ ) || defined( __ppc__ ) ) I'm using VxWorks 6.5. Their version of gcc defines __ppc but not __ppc__. This is some of the output from "ccppc -dD -E emptyfile.cpp": #define __GNUC__ 3 #define _ARCH_PPC 1 #define __ppc 1 #define CPU_FAMILY PPC #define CPU PPC604 #define __vxworks 1 #define __VXWORKS__ 1 If the sp_counted_base.hpp check was written like this: #elif defined( __GNUC__ ) && ( defined( __powerpc__ ) || defined( __ppc__ ) || defined( __ppc ) ) Then I would see the correct file get included. Is this a change that should be made to the boost source or do you feel it is a problem with my compiler? Regards, -Dan

2 1

[Serialization] Getting link errors in gcc 4.1.2 with boost serialization library 1.3.4
by Reusser, Edward 20 Sep '07

20 Sep '07

I am using the boost serialization library version 1.3.4. This compiles fine. However, when I link it into my code, I am getting the following link errors: /home1/ken/main_gcc41/bld/sdpm/linobj/libsdpm.so: undefined reference to `boost::archive::basic_binary_oprimitive<boost::archive::binary_oarchive , _Dinkum_std::basic_ostream<char, _Dinkum_std::char_traits<char> > >::~basic_binary_oprimitive()' /home1/ken/main_gcc41/bld/sdpm/linobj/libsdpm.so: undefined reference to `boost::archive::basic_binary_oprimitive<boost::archive::binary_oarchive , _Dinkum_std::basic_ostream<char, _Dinkum_std::char_traits<char> > >::basic_binary_oprimitive(_Dinkum_std::basic_ostream<char, _Dinkum_std::char_traits<char> >&, bool)' /home1/ken/main_gcc41/bld/sdpm/linobj/libsdpm.so: undefined reference to `boost::archive::basic_binary_iprimitive<boost::archive::binary_iarchive , _Dinkum_std::basic_istream<char, _Dinkum_std::char_traits<char> > >::load(_Dinkum_std::basic_string<char, _Dinkum_std::char_traits<char>, _Dinkum_std::allocator<char> >&)' /home1/ken/main_gcc41/bld/sdpm/linobj/libsdpm.so: undefined reference to `boost::archive::basic_binary_oprimitive<boost::archive::binary_oarchive , _Dinkum_std::basic_ostream<char, _Dinkum_std::char_traits<char> > >::save(_Dinkum_std::basic_string<char, _Dinkum_std::char_traits<char>, _Dinkum_std::allocator<char> > const&)' /home1/ken/main_gcc41/bld/sdpm/linobj/libsdpm.so: undefined reference to `boost::archive::basic_binary_iprimitive<boost::archive::binary_iarchive , _Dinkum_std::basic_istream<char, _Dinkum_std::char_traits<char> > >::~basic_binary_iprimitive()' /home1/ken/main_gcc41/bld/sdpm/linobj/libsdpm.so: undefined reference to `boost::archive::basic_binary_iprimitive<boost::archive::binary_iarchive , _Dinkum_std::basic_istream<char, _Dinkum_std::char_traits<char> > >::basic_binary_iprimitive(_Dinkum_std::basic_istream<char, _Dinkum_std::char_traits<char> >&, bool)' I am not seeing these using previous versions of the compiler. Ed The information contained in or attached to this email may be subject to the Export Administration Regulations (EAR), administered by the U.S. Department of Commerce, or the International Traffic in Arms Regulations (ITAR), administered by the U.S. Department of State, and may require an export license from the Commerce or State Department prior to its export. An export can include a release or disclosure to a foreign national inside or outside the United States. Include this notice with any reproduced portion of this information.

1 0

[multi_index] composite_key question
by Jeffrey Holle 20 Sep '07

20 Sep '07

I've a multi_index container that has a composite_key index which is used to express a "checker board" like concept. I need to iterator over all rows, from 0 to n. I know how to employ equal_range(make_tuple(y)) to get the elements within a specific row, but how do I determine n? The only way I can think of doing this is doing a linear search using another index. Is there a better way?

3 2

how to use map_list
by swapna 20 Sep '07

20 Sep '07

hi, Does anone know how to use map_list_of and the BOOST_foreach with the created map. Thanks Swapna --------------------------------- Forgot the famous last words? Access your message archive online. Click here.

2 1

Re: [Boost-users] polymorphism and boost::shared_ptr
by Daniel P Furlani 20 Sep '07

20 Sep '07

On 9/7/07, Brad Ryder <bradryder67_at_[hidden]> wrote: > typedef boost::shared_ptr<A> shPtrA; > typedef boost::shared_ptr<B> shPtrB; > typedef boost::shared_ptr<C> shPtrC; > > class A > { > > void AddChild(shPtrA& child); > private: > std::vector<shPtrA> m_children; > } > > class B : public A > { > > } > > class C: public B > { > > } > > main() > { > A base; > shPtrB spB(new B()); > shPtrC spC(new C()); > > // I dont want to typecast spB or spC > > A.AddChild( (shPtrA) spB); > A.AddChild( (shPtrA) spC); > } Hello, I am also new to the list, and saw your question in the archives. You don't need to use static_pointer_cast<> or limit yourself to using base class shared_ptr<>s, as has been suggested. The solution is to pass by value instead of by reference. Then the compiler has a chance to perform an implicit conversion from one shared_ptr type to another. In other words, just remove the "&" from the signature of AddChild(). void AddChild(shPtrA child); The conversion won't occur when you pass by reference because shPtrB does not inherit from shPtrA. HTH -Dan

1 0

[serialization] Error with bitfields under PathScale EKO compiler
by Bill Lear 20 Sep '07

20 Sep '07

I have a standard serialize method, working under various flavors of gcc and the Intel C++ compilers: struct ElementEnum { enum Origin { Standard, Base, Custom }; }; class Foo { ... ElementEnum::Origin origin : 4; template<...> void serialize(...) { ar & BOOST_SERIALIZATION_NVP(origin); } }; However, under the PathScale EKO compiler (pathCC), I get an error: "error: invalid initialization of reference of type Origin& from expression of type const unsigned char:4." Apparently make_nvp takes as its second argument a non-const reference. It appears that non-const references are illegal under this compiler, and, who knows, perhaps the latest C++ standard?? I was able to create a test case that caused this same error. However, the test case also failed under gcc and the Intel compiler, so the test case (appended below after my sig) is not narrow enough, apparently. If I rewrite the serialization method to not use the BOOST_SERIALIZATION_NVP macro at all, the code works. Any help on working around this (aside from tearing out BOOST_SERIALIAZATION_NVP() calls altogether) would be appreciated. Bill // test case to provoke the same error that I got compiling // serialization code under PathScale EKO compiler. #include <iostream> using namespace std; template <class T> void print(T& t) { cout << "t is now: " << t << endl; } struct Foo { int itsie : 4; }; int main() { Foo f; print(f.itsie); }

2 3

boost::ptr_vector null check bloat
by Miles Bader 20 Sep '07

20 Sep '07

I tried changing some of my code that used std::vector<foo *> to using boost::ptr_vector<foo> (to get automatic deallocation and the slightly nicer syntax), but the resulting compiled code seems oddly bloated compared to the original code using std::vector. My intuition is that the two should generally compile to largely equivalent object code, except where there's a necessary functional difference. Some of the problem seems to be that it does null-pointer checking at _access_ time, instead of when adding new entries. The result is that (for instance) calls to operator[] end up with an "unnecessary" null check, which drags in tons of stuff related to the accompanying assertion failure, some of which is apparently unshared across object files, like the long __PRETTY_FUNCTION__ strings passed to the assertion failure call. What's the general impression people have of boost::ptr_vector and related classes? Do people use it? Does it have a "bloat" reputation? Thanks, -Miles -- "Suppose we've chosen the wrong god. Every time we go to church we're just making him madder and madder." -- Homer Simpson

4 8