- Boost - lists.boost.org

spirit typeof tests failing for cw-8_3
by Arkadiy Vertleyb 25 Feb '06

25 Feb '06

I am not currently planning to support emulation mode for this compiler. AFAIU, there are no plans to support this compiler beyond 1.34, so I figured it doesn't worth the effort (and extra code for workarounds), especially since the native mode is there. Please let me know if having emulation mode on cw-8_3 is important for some reason. Regards, Arkadiy

1 0

[serialization] BOOST_CLASS_EXPORT in a separate static library
by Marcin Kalicinski 24 Feb '06

24 Feb '06

I noticed strange behavior of serialization on VC 8.0. I use BOOST_CLASS_EXPORT macro in a .cpp file to register a derived class, so that it can be properly serialized through base class pointer. This works fine, as long the .obj file resulting from compilation of .cpp file which uses BOOST_CLASS_EXPORT macro is linked directly to the final executable. But when I first put this .obj file into a static library, and then link the library to executable, I start getting unregistered_class exceptions - as if there was no BOOST_CLASS_EXPORT invoked. So far I believed that there is no difference between linking to object file and to static library file containing that object file. It seems my beliefs were now shattered. Has anybody else had this issue before? Thank you, Marcin

3 2

[bind][function_types] Why do bind objects not have an ::arity, ... interface ?
by Peter Soetens 24 Feb '06

24 Feb '06

Hi, I want to use a partial template specialisation which is selected on basis of the arity of a bind function object type. Event more, I'd like to access the ::arg1_type, etc types of a bind object type. Why is this not part of the boost::bind or lambda::bind interface ? I thought solving it with the (to be accepted) function_types::function_type_arity<BoundType>::type::value, but the latest version of function_types does not 'support' inspection of bound object types. Other library writers must have felt a need for this interface as well ? Or am I missing something ? With kind regards, Peter Soetens -- Peter Soetens -- FMTC -- <http://www.fmtc.be>

3 2

Re: [boost] [serialization] [multi_index][cw-8_3] help request forregressions in CW 8.3 Windows
by JOAQUIN LOPEZ MU?Z 24 Feb '06

24 Feb '06

----- Mensaje original ----- De: David Abrahams <dave(a)boost-consulting.com> Fecha: Viernes, Febrero 24, 2006 5:38 pm Asunto: [boost] [serialization] [multi_index][cw-8_3] help request forregressions in CW 8.3 Windows > Joaquín Mª López Muñoz <joaquin(a)tid.es> writes: > > > Excuse my insistence, but I have no other way to try to fix current > > regressions in Boost.MultiIndex but to ask for external support. So > > I'm reposting my plea for help: > > > > Hello, > > > > Most current tests for Boost.MultiIndex under CW 8.3 in Windows > > fail at run time with the following: > > > > Running 1 test case... > > unknown location(0): fatal error in "test_main_caller( argc, > argv )": > > unknown type > > > > *** 1 failure detected in test suite "Test Program" > > > > EXIT STATUS: 201 > > > > (see for instance http://tinyurl.com/fqarz) compilation and > building> having gone fine. As this is giving very little info, > I'd be grateful if > > > > someone with local access to CW 8.3 Windows (it is likely that > > Mac will also do) volunteer to do some exploration of the > > problem. If you are such volunteer, please contact me privately. > > I tried it; I'm mostly seeing this: > > cw.compile.c++ ..\..\..\bin.v2\libs\serialization\build\cw- > 8.3\debug\basic_serializer_map.obj ### mwcc Compiler: > # In: ..\..\..\boost\archive\detail\basic_serializer_map.hpp > # From: ..\..\serialization\src\basic_serializer_map.cpp > # --------------------------------------------------------- > # 56: }; > # Error: ^ > # using implicit copy assigment for class with const or > reference member ('boost::archive::detail::basic_serializer_map') > > Errors caused tool to abort. > > which is a problem in serialization. Aside from that all the tests > are passing for me with cw-8. Do you mean that multi_index/test_basic, for instance, passes? What could be then the reason for the discrepancy with http://tinyurl.com/fqarz? Are there maybe upgrades to the CW 8.3 compiler and your version differs from that of metacomm? Thanks for your input, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

2 1

[rerepost][multi_index][cw-8_3] help request for regressions in CW 8.3 Windows
by Joaquín Mª López Muñoz 24 Feb '06

24 Feb '06

Excuse my insistence, but I have no other way to try to fix current regressions in Boost.MultiIndex but to ask for external support. So I'm reposting my plea for help: Hello, Most current tests for Boost.MultiIndex under CW 8.3 in Windows fail at run time with the following: Running 1 test case... unknown location(0): fatal error in "test_main_caller( argc, argv )": unknown type *** 1 failure detected in test suite "Test Program" EXIT STATUS: 201 (see for instance http://tinyurl.com/fqarz) compilation and building having gone fine. As this is giving very little info, I'd be grateful if someone with local access to CW 8.3 Windows (it is likely that Mac will also do) volunteer to do some exploration of the problem. If you are such volunteer, please contact me privately. Thank you very much, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

2 1

[multi_index] key swap status
by gruenedd＠idmt.fraunhofer.de 24 Feb '06

24 Feb '06

Hello, What is the status of key swapping, issued here[1]? Do you see a chance to have this included sometime? :] Best Regards, David [1] http://lists.boost.org/boost-users/2005/04/11223.php ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.

3 2

boost::null
by Raffaele Romito 24 Feb '06

24 Feb '06

What about a null_helper class? This is my approach: namespace boost { static class null_helper { public: template <typename T> inline operator shared_ptr<T>() const { return shared_ptr<T>(); } template <typename T> inline operator T *() const { return 0; } }null; } I find this very useful in samples like this: class node { public: node(boost::shared_ptr<node> parent = boost::null, // instead of boost::shared_ptr<node>() int *value = boost::null); // no difference from ordinary pointers }; Waiting for your feedback ;) ------------- Raffaele Romito

1 0

[typeof] help needed with gcc 3.2.x/3.3.x linux
by Arkadiy Vertleyb 23 Feb '06

23 Feb '06

Could anybody with one of those compilers installed help me figure out why they complain about this code: //--------------- #include <boost/typeof/typeof.hpp> #include <boost/type_traits/is_same.hpp> #include <boost/static_assert.hpp> void f() {} template<class T> class x { BOOST_STATIC_ASSERT(( boost::is_same<BOOST_TYPEOF_TPL(&f), void(*)()>::value )); }; //--------------- I suspect the problem is related to the test itself rather than to typeof, and might be workarounded (or is it "worked around"?) with a couple of intermediate typedefs, but I might be wrong... Thanks in advance, Arkadiy

2 2

[detail] utf8 patch for Borland 2006
by AlisdairM 23 Feb '06

23 Feb '06

Not sure who is maintainer of this file? Anyway, following patch is a little rough, not testing on compiler version, but is required for the current Borland compiler and the Dinkumware libraries it ships with. Fixing this should fix tests across a few libraries: cvs diff -u -wb -- boost\detail\utf8_codecvt_facet.hpp (in directory E:\sourceforge\devel\boost\) Index: boost/detail/utf8_codecvt_facet.hpp =================================================================== RCS file: /cvsroot/boost/boost/boost/detail/utf8_codecvt_facet.hpp,v retrieving revision 1.3 diff -u -w -b -r1.3 utf8_codecvt_facet.hpp --- boost/detail/utf8_codecvt_facet.hpp 20 Jan 2006 16:53:30 -0000 1.3 +++ boost/detail/utf8_codecvt_facet.hpp 21 Jan 2006 19:29:24 -0000 @@ -92,7 +92,7 @@ namespace std { #if defined(__LIBCOMO__) using ::mbstate_t; - #elif defined(BOOST_DINKUMWARE_STDLIB) + #elif defined(BOOST_DINKUMWARE_STDLIB) && !defined(__BORLANDC__) using ::mbstate_t; #elif defined(__SGI_STL_PORT) #elif defined(BOOST_NO_STDC_NAMESPACE) ***** CVS exited normally with code 1 ***** -- AlisdairM

2 2

[smart_ptr] poly casts
by David Gruener 23 Feb '06

23 Feb '06

Hello, It seems that the polymorphic casts are deprecated without replacement with the new *_pointer_cast style. Why? David.

4 3