January 2009 - Boost - lists.boost.org

BoostCon '09 Hotel Accepting Reservations
by David Abrahams 09 Jan '09

09 Jan '09

Hello Boosters and former BoostCon attendees! The BoostCon website has been updated for 2009, and we expect to open registration in just a few days. However, the conference hotel is already prepared to start taking our reservations. If you are planning to go to Boostcon this year, I suggest you take a few minutes to make your reservation today. Why you should book your room now: http://www.boostcon.com/news/2009/Jan/01/hotel-now-accepting-reservations Go straight to reservations: http://tinyurl.com/boostcon09hotel Cheers! -- Dave Abrahams Conference Chair http://boostcon.com

1 0

[bind] virtual function pointer woes
by Eric Niebler 09 Jan '09

09 Jan '09

Imagine I have a base class B with a virtual function f() that is overridden in a derived class D ... struct B { virtual void f() { std::printf("B\n"); } }; struct D : B { virtual void f() { std::printf("D\n"); } }; Now, I can have a derived object D and call the base member directly using qualification: D d; d.B::f(); // OK, called B::f But if I take the address of B::f for use with boost::bind, I get a thunk that always uses virtual dispatch... boost::bind( &B::f, _1 )( d ); // Hrm, calls D::f What if I don't want the virtual dispatch, and I really want this to behave like d.B::f()? I know this is not specifically a problem with boost::bind, but a more general property of (virtual) member function pointers. Is there any way to get a non-virtual member function pointer to a virtual member function? If there is a way, I'm sure I don't know it. -- Eric Niebler BoostPro Computing http://www.boostpro.com

3 4

[1.38.0] Deadlines looming
by Beman Dawes 09 Jan '09

09 Jan '09

The 1.38.0 beta is scheduled for a little more than a week from now. Release branch test reporting has been stalled, but we are working to get reporting restarted so the beta can go out more or less on time. Please complete any needed merges by Friday, January 16th. Thanks, --Beman

1 0

ASL patches
by Michael Marcin 09 Jan '09

09 Jan '09

I was playing with Adam and Eve and noticed that Adobe has some patches against boost 1.37, should any of these be merged into trunk to be included in 1.38 that haven't been? There looks to be patches for: /boost/detail/is_incrementable.hpp /boost/test/floating_point_comparison.hpp /tools/build/v2/tools/darwin.jam /tools/build/v2/tools/msvc.jam /tools/build/v2/tools/testing.jam /tools/jam/src/expand.c -- Michael Marcin diff -wur --unidirectional-new-file -x status -x adobe_information -xp4config -x bin.v2 ../diffable_boost_1_37_0/boost/detail/is_incrementable.hpp ../boost_libraries/boost/detail/is_incrementable.hpp --- ../diffable_boost_1_37_0/boost/detail/is_incrementable.hpp 2008-03-22 14:45:55.000000000 -0700 +++ ../boost_libraries/boost/detail/is_incrementable.hpp 2008-11-14 20:42:11.687500000 -0800 @@ -65,10 +65,10 @@ # endif // two check overloads help us identify which operator++ was picked - char (& check(tag) )[2]; + char (& (check)(tag) )[2]; template <class T> - char check(T const&); + char (check)(T const&); template <class T> @@ -78,7 +78,7 @@ BOOST_STATIC_CONSTANT( bool - , value = sizeof(is_incrementable_::check(BOOST_comma(++x,0))) == 1 + , value = sizeof((is_incrementable_::check)(BOOST_comma(++x,0))) == 1 ); }; @@ -89,7 +89,7 @@ BOOST_STATIC_CONSTANT( bool - , value = sizeof(is_incrementable_::check(BOOST_comma(x++,0))) == 1 + , value = sizeof((is_incrementable_::check)(BOOST_comma(x++,0))) == 1 ); }; } diff -wur --unidirectional-new-file -x status -x adobe_information -xp4config -x bin.v2 ../diffable_boost_1_37_0/boost/test/floating_point_comparison.hpp ../boost_libraries/boost/test/floating_point_comparison.hpp --- ../diffable_boost_1_37_0/boost/test/floating_point_comparison.hpp 2008-10-13 01:20:26.000000000 -0700 +++ ../boost_libraries/boost/test/floating_point_comparison.hpp 2008-11-05 17:53:37.546875000 -0800 @@ -220,8 +220,16 @@ } }; +template<class T> +struct static_ { + static T value; +}; + +template<class T> +T static_<T>::value; + namespace { -check_is_close_t check_is_close; + const check_is_close_t& x = static_<check_is_close_t>::value; } //____________________________________________________________________________// @@ -243,7 +251,7 @@ }; namespace { -check_is_small_t check_is_small; + const check_is_small_t& y = static_<check_is_small_t>::value; } //____________________________________________________________________________// diff -wur --unidirectional-new-file -x status -x adobe_information -xp4config -x bin.v2 ../diffable_boost_1_37_0/tools/build/v2/tools/darwin.jam ../boost_libraries/tools/build/v2/tools/darwin.jam --- ../diffable_boost_1_37_0/tools/build/v2/tools/darwin.jam 2008-10-16 01:42:03.000000000 -0700 +++ ../boost_libraries/tools/build/v2/tools/darwin.jam 2008-11-05 14:07:26.656250000 -0800 @@ -147,21 +147,22 @@ flags darwin.link OPTIONS $(condition)/<main-target-type>EXE/<address-model>/<debug-symbols>off : -s ; } - else - { - # Otherwise we need to find a strip program to use. And hence - # also tell the link action that we need to use a strip - # post-process. - flags darwin.link NEED_STRIP $(condition)/<debug-symbols>off : "" ; - strip = - [ common.get-invocation-command darwin - : strip : [ feature.get-values <striper> : $(options) ] : $(bin) : search-path ] ; - flags darwin.link .STRIP $(condition) : $(strip[1]) ; - if $(.debug-configuration) - { - ECHO notice: using strip :: $(condition) :: $(strip[1]) ; - } - } +## produces bad code on macs (release). maybe different options need to be sent to strip +# else +# { +# # Otherwise we need to find a strip program to use. And hence +# # also tell the link action that we need to use a strip +# # post-process. +# flags darwin.link NEED_STRIP $(condition)/<debug-symbols>off : "" ; +# strip = +# [ common.get-invocation-command darwin +# : strip : [ feature.get-values <striper> : $(options) ] : $(bin) : search-path ] ; +# flags darwin.link .STRIP $(condition) : $(strip[1]) ; +# if $(.debug-configuration) +# { +# ECHO notice: using strip :: $(condition) :: $(strip[1]) ; +# } +# } # - The archive builder (libtool is the default as creating # archives in darwin is complicated. diff -wur --unidirectional-new-file -x status -x adobe_information -xp4config -x bin.v2 ../diffable_boost_1_37_0/tools/build/v2/tools/msvc.jam ../boost_libraries/tools/build/v2/tools/msvc.jam --- ../diffable_boost_1_37_0/tools/build/v2/tools/msvc.jam 2008-10-16 01:42:03.000000000 -0700 +++ ../boost_libraries/tools/build/v2/tools/msvc.jam 2008-11-05 14:07:25.046875000 -0800 @@ -973,6 +973,9 @@ # Declare msvc toolset specific features. { + # Secure STL feature + feature.feature stl-security-theater : off on : propagated link-incompatible ; + feature.feature debug-store : object database : propagated ; feature.feature pch-source : : dependency free ; } @@ -1040,6 +1043,8 @@ toolset.flags msvc.compile C++FLAGS <exception-handling>on/<asynch-exceptions>off/<extern-c-nothrow>on : /EHsc ; toolset.flags msvc.compile C++FLAGS <exception-handling>on/<asynch-exceptions>on/<extern-c-nothrow>off : /EHa ; toolset.flags msvc.compile C++FLAGS <exception-handling>on/<asynch-exceptions>on/<extern-c-nothrow>on : /EHac ; + toolset.flags msvc.compile C++FLAGS <stl-security-theater>off : /D_SECURE_SCL=0 ; + toolset.flags msvc.compile C++FLAGS <stl-security-theater>on : /D_SECURE_SCL=1 ; # By default 8.0 enables rtti support while prior versions disabled it. We # simply enable or disable it explicitly so we do not have to depend on this diff -wur --unidirectional-new-file -x status -x adobe_information -xp4config -x bin.v2 ../diffable_boost_1_37_0/tools/build/v2/tools/testing.jam ../boost_libraries/tools/build/v2/tools/testing.jam --- ../diffable_boost_1_37_0/tools/build/v2/tools/testing.jam 2008-10-16 01:42:03.000000000 -0700 +++ ../boost_libraries/tools/build/v2/tools/testing.jam 2008-11-13 09:27:41.453125000 -0800 @@ -62,6 +62,7 @@ feature.feature testing.arg : : free incidental ; feature.feature testing.input-file : : free dependency ; +feature.feature preserve-test-targets : off on : incidental propagated ; # Register target types. type.register TEST : test ; @@ -393,11 +394,6 @@ local argv = [ modules.peek : ARGV ] ; -if --preserve-test-targets in $(argv) -{ - preserve-test-targets = true ; -} - toolset.flags testing.capture-output ARGS <testing.arg> ; toolset.flags testing.capture-output INPUT_FILES <testing.input-file> ; @@ -440,7 +436,7 @@ run-path-setup $(target) : $(source) : $(properties) ; - if ! $(preserve-test-targets) + if [ feature.get-values preserve-test-targets : $(properties) ] = off { TEMPORARY $(targets-to-remove) ; # Set a second action on target that will be executed after capture diff -wur --unidirectional-new-file -x status -x adobe_information -xp4config -x bin.v2 ../diffable_boost_1_37_0/tools/jam/src/expand.c ../boost_libraries/tools/jam/src/expand.c --- ../diffable_boost_1_37_0/tools/jam/src/expand.c 2008-07-29 04:43:31.000000000 -0700 +++ ../boost_libraries/tools/jam/src/expand.c 2008-11-05 14:07:25.984375000 -0800 @@ -749,7 +749,7 @@ assert(l != 0); assert(list_next(l) == 0); # ifdef OS_CYGWIN - assert( !strcmp( l->string, "c:\\foo\\bar" ) ); + assert( strcmp( l->string, "c:\\foo\\bar" ) == 0 || strcmp( l->string, "C:\\foo\\bar" ) == 0 ); # else assert( !strcmp( l->string, cygpath ) ); # endif

2 1

Testing -fno-rtti builds (was: [function] function_base.hpp:220: error: cannot use typeid with -fno-rtti)
by Emil Dotchevski 08 Jan '09

08 Jan '09

On Thu, Jan 8, 2009 at 9:21 AM, Doug Gregor <doug.gregor(a)gmail.com> wrote: > On Wed, Jan 7, 2009 at 5:33 PM, Steven Watanabe <watanabesj(a)gmail.com> wrote: >> Emil Dotchevski wrote: >>> >>> Does boost::function not support -fno-rtti by design (linux/gcc 4.1.3)? >> >> target, target_type, and contains require typeid. I suppose that in >> principle these could be disabled. > > Yes, it's possible, and I'd certainly accept a patch that did so. Someone correct me if I'm doing it wrong, but for example the RTTI-less and typeid-less configurations of Boost Exception, the library I'm maintaining, are tested only by me on my own computer, and I'm only testing them with MSVC and a single version of linux/gcc. Shouldn't the automatic trunk/release tests handle these, as well as other supported configurations? I guess what I'm saying is, I can make a patch to make boost::function work without RTTI, but I'm certainly not up to running those tests too. :) -- Emil Dotchevski Reverge Studios, Inc. http://www.revergestudios.com/reblog/index.php?n=ReCode

3 6

[function] function_base.hpp:220: error: cannot use typeid with -fno-rtti
by Emil Dotchevski 08 Jan '09

08 Jan '09

Does boost::function not support -fno-rtti by design (linux/gcc 4.1.3)? -- Emil Dotchevski Reverge Studios, Inc. http://www.revergestudios.com/reblog/index.php?n=ReCode

3 2

[lambda] patch to make lambda recognize operator+ for std::string
by Steven Watanabe 08 Jan '09

08 Jan '09

AMDG The attached patch fixes https://svn.boost.org/trac/boost/ticket/781 Okay to commit? In Christ, Steven Watanabe

2 1

Detecting if a class overrides free store functions
by Michael Marcin 08 Jan '09

08 Jan '09

I want to create a STL allocator that uses a class's static new function if it exists and will fallback on std::allocator::allocate if it doesn't. I think it should be easy to do this with sfinae but I don't see an appropriate type trait to detect if a class provides allocation and deallocation functions. Should these be provided? Maybe boost::has_allocation_function<T> boost::has_deallocation_function<T> or just boost::has_new<T> boost::has_delete<T> -- Michael Marcin

3 3

[smart ptr] implicit auto_ptr r-value initialization for scoped_ptr?
by Jason Moore 08 Jan '09

08 Jan '09

This is absolute genius!!! (i.e. implicit only for r-values of type std::auto_ptr<T>) std::auto_ptr<T> Create(); ... if (boost::shared_ptr<T> my_ptr = Create()) { my_ptr->doIt(); } Why is this feature not provided for boost::scoped_ptr? Thanks, Jason Moore

1 0

is_convertible problems associated with move-only types
by Howard Hinnant 07 Jan '09

07 Jan '09

I'm continuing to refine a boost::unique_ptr emulation and testsuite. One thing I've just uncovered is that I'm hitting a brick wall with is_convertible and move-only deleters used within the unique_ptr move constructor. The problem is that boost::is_convertible<From, To> assumes From is an lvalue. is_convertible<Deleter, Deleter> gets instantiated and it is all over. I've modified my copy of is_convertible (without really knowing what I'm doing) like so: Index: is_convertible.hpp =================================================================== --- is_convertible.hpp (revision 50433) +++ is_convertible.hpp (working copy) @@ -119,6 +119,8 @@ struct any_conversion { template <typename T> any_conversion(const volatile T&); + template <typename T> any_conversion(volatile T&); + template <typename T> any_conversion(const T&); template <typename T> any_conversion(T&); }; @@ -131,8 +133,8 @@ template <typename From, typename To> struct is_convertible_basic_impl { - static From _m_from; - static bool const value = sizeof( detail::checker<To>::_m_check(_m_from, 0) ) + static From _m_from(); + static bool const value = sizeof( detail::checker<To>::_m_check(_m_from(), 0) ) == sizeof(::boost::type_traits::yes_type); }; @@ -291,7 +293,8 @@ template <typename From, typename To> struct is_convertible_impl { - typedef typename add_reference<From>::type ref_type; +// typedef typename add_reference<From>::type ref_type; + typedef From ref_type; BOOST_STATIC_CONSTANT(bool, value = (::boost::type_traits::ice_and< ::boost::type_traits::ice_or< This hits just the part targeting gcc. The intent is that in boost::is_convertible<From, To> From is now considered an rvalue (unless From is a (lvalue) reference type). This makes my test case happy. I can now move construct a unique_ptr with a move-only deleter. Fwiw, this definition of is_convertible is consistent with std::is_convertible in N2800: > In order to instantiate the template is_convertible<From, To>, the > following code shall be well formed: > > template <class T> > typename add_rvalue_reference<T>::type create(); > To test() { > return create<From>(); > } > > [ Note: This requirement gives well defined results for reference > types, > void types, array types, and function types. -- end note ] I've got a much more refined unique_ptr testsuite on the way which will contain the unique_ptr header and test case for this (not there yet, I'll post when it is). -Howard

3 9