- Boost - lists.boost.org

[optional] swapping initialized with uninitialized
by Vaclav Vesely 11 Mar '06

11 Mar '06

Hi, When two optionals (one initialized and other not) are swaped, then the one value in uninitialized optional is constructed and the other one in initialized optional is destructed. IMHO it's not necessary. It should be more effective to swap optional's m_storages without any values constructiona and destruction. Is there any reason for the current implementation? Regards, Vaclav

3 2

boost::serialization - problem move to 1.33.1 version
by Алексей Обухов 11 Mar '06

11 Mar '06

Hello all ! I write small test program using boost::serialization. If I use boost 1.32 - compilation OK. But I change boost to 1.33.1 - compiler rise error :( Where I be mistaken ? Compiler messages: OtherTest.cpp w:\Devil\Projects\PLTools2_boost_1.33.1\boost\boost\serialization\export.hpp(146) : error C2027: use of undefined type 'boost::serialization::extended_type_info_null' with [ T=BoostTest2::A ] w:\Devil\Projects\PLTools2_boost_1.33.1\boost\boost\serialization\export.hpp(146) : see reference to class template instantiation 'boost::serialization::extended_type_info_null' being compiled with [ T=BoostTest2::A ] w:\Devil\Projects\PLTools2_boost_1.33.1\boost\boost\serialization\export.hpp(145) : while compiling class-template member function 'void boost::archive::detail::guid_initializer::export_register(const char *)' with [ T=BoostTest2::A ] w:\Devil\Projects\PLTools2_boost_1.33.1\Tests\OtherTest\boost_test_2.h(64) : see reference to class template instantiation 'boost::archive::detail::guid_initializer' being compiled with [ T=BoostTest2::A ] Results This error point to line: (full code below) BOOST_CLASS_EXPORT_GUID( BoostTest2::B, "ClassB" ) OtherTest.cpp: #include <iostream> #include <sstream> #include <boost/serialization/vector.hpp> #include <boost/serialization/string.hpp> #include <boost/serialization/nvp.hpp> #include <boost/archive/xml_iarchive.hpp> #include <boost/archive/xml_oarchive.hpp> #include <boost/serialization/export.hpp> #include <boost/serialization/base_object.hpp> #include "boost_test_2.h" using namespace BoostTest2; /// ..... boost_test_2.h #pragma once namespace BoostTest2 { class A { public: A() {m_A = 1;} virtual ~A() { } int m_A; }; class B : public A { public: B() {m_B = 2;} virtual ~B() {} int m_B; }; class C : public A { public: C() {m_C = 3;} virtual ~C() {} int m_C; }; } #define BOOST_NVP(name,value) \ boost::serialization::make_nvp(name, value) namespace boost { namespace serialization { template<class Archive> void serialize(Archive & ar, BoostTest2::A &a, const unsigned int version) { ar & BOOST_NVP( "A", a.m_A ); } template<class Archive> void serialize(Archive & ar, BoostTest2::B &b, const unsigned int version) { ar & boost::serialization::make_nvp( "A", boost::serialization::base_object<A>(b) ); ar & BOOST_NVP( "B", b.m_B ); } template<class Archive> void serialize(Archive & ar, BoostTest2::C &c, const unsigned int version) { ar & boost::serialization::make_nvp( "A", boost::serialization::base_object<A>(c) ); ar & BOOST_NVP( "C", c.m_C ); } } } BOOST_CLASS_EXPORT_GUID( BoostTest2::B, "ClassB" ) BOOST_CLASS_EXPORT_GUID( BoostTest2::C, "ClassC" ) BOOST_IS_ABSTRACT( BoostTest2::A ) BOOST_CLASS_VERSION( BoostTest2::A , 0 )

1 0

Re: [boost] [interprocess] shared memory lifetime
by Olaf van der Spek 10 Mar '06

10 Mar '06

On 3/10/06, Kim Barrett <kab(a)irobot.com> wrote: > >One option is to use a memory mapped file to emulate posix-like behavior > >in windows. But it surely will be slower (the OS will dump data to file) > >and the unlink is not trivial. > > That does seem like it would be bad if true. Are there any file > manipulation options that might be used to control that? I don't know, > but I'll ask around and do some reading and see what I can find. > > [Later] As long as you don't flush the file it would seem like a memory > mapped file should be no different from a shared memory in performance, > the only difference is where data gets written if it needs to be swapped > out. The big downside to using an actual file would be that you need a > place in the file system to put it. Then you get into the whole name of > the shared memory portability problem again. The OS will probably write dirty pages to disk even before paging them out so that if it needs to page them out it doesn't have to also do a disk write.

1 0

[Announce] Mailing list downtime this Saturday
by Doug Gregor 10 Mar '06

10 Mar '06

All Boost mailing lists hosted at lists.boost.org will be down from 8pm to midnight EST this Saturday, March 11th while we migrate to a new, dedicated mail server provided by the Open Systems Lab at Indiana University. During this down time, the mail archives at lists.boost.org will be unavailable and any mail sent to lists.boost.org will be queued and forwarded. We ask that members refrain from posting messages to lists.boost.org during this time "just in case," although we expect that all mail sent during this time will be forwarded properly. If you should have any questions, please contact me directly at dgregor(a)cs.indiana.edu. Doug Gregor Boost Moderator dgregor(a)cs.indiana.edu

1 0

[FileSystem] Incorrect info returned by filesystem::filesystem_error exception
by james＠massey46.fsnet.co.uk 10 Mar '06

10 Mar '06

Hi, There appears to be problems with boost::filesystem::rename with respect to the information returned by boost:filesystem::filesystem_error exception. See the 2 tests below and attached source code examples. Test 1: Precondition: The "Source" file does not exist and the "Destination" file does exist. Test: Perform rename(Source,Destination); Expected Result: The rename fails because the Source file does not exist. Actual Result: A boost filesystem exception is thrown and the output is what text: boost::filesystem::rename: "Source", "Destination": Success error_code: 1 native_error: 0 The exception information suggests the command was successful. The result of a UNIX "mv" with the same scenario is "mv: cannot stat `Source': No such file or directory" Note that a similar output is displayed if the "Source" and "Destination" files exist before the test is run. Test 2: Precondition: The "Source" and "Destination" files exist. The directory DirWithNoWritePermission exists and has permissions 555. Tests: Perform rename(Source,DirWithNoWritePermission/Anything); Perform rename(Source,Destination); Expected Result: The first rename fails because the directory permissions forbid the rename. The second rename fails because the destination file exists. Actual Result: A boost filesystem exception is thrown and the output is what text: boost::filesystem::rename: "Source", "DirWithNoWritePermission/Anything": Permission denied error_code: 3 native_error: 13 and is the same in both cases. It appears that the exception information displayed for the second test has been "recalled" from the first test. This may be related to the problem shown in Test 1. Test programs run under Suse Enterprise Linux 9 with gcc version 3.3.3, boost 1_33_0 and 1_33_1 Regards James Massey

1 0

[statechart] - error with templated states (no type 'inner_contex_type')
by Oliver.Kowalke＠infineon.com 10 Mar '06

10 Mar '06

Hello, I get another error if I use templated states (gcc-3.4.6). 'simple_state.hpp:403: error: no type named inner_context_type in struct proto<int>' Instantiation of state_machine:: initiate() causes the error. with regards, Oliver #include <iostream> #include <stdexcept> #include <boost/statechart/simple_state.hpp> #include <boost/statechart/state_machine.hpp> namespace sc = boost::statechart; template< typename S > struct proto; template< typename S > struct sm : public sc::state_machine< sm< S >, proto< S > > {}; template< typename S > struct inactive; template< typename S > struct proto : public sc::simple_state< proto< S >, sm< S >, inactive< S > > {}; template< typename S > struct inactive : public sc::simple_state< inactive< S >, proto< S > > {}; int main() { try { sm< int > fsm; fsm.initiate(); // <<== causes compiler error return 0; } catch ( std::runtime_error const& e) { std::cerr << e.what() << std::endl; } catch (...) { std::cerr << "unhandled exception" << std::endl; } return -1; }

3 3

[optional] swapping initialized with uninitialized
by Vaclav Vesely 10 Mar '06

10 Mar '06

Hi, When two optionals (one initialized and other not) are swaped, then the one value in uninitialized optional is constructed and the other one in initialized optional is destructed. IMHO it's not necessary. It should be more effective to swap optional's m_storages without any values constructiona and destruction. Is there any reason for the current implementation? Regards, Vaclav

1 0

in_place_factory0
by Vaclav Vesely 10 Mar '06

10 Mar '06

Hi, is there any reason for missing in_place_factory without arguments? I would like to write optional<my_class> mc(in_place()); I've wrote the code for this. But I can't modify the macro hell in "in_place_factory.hpp". Regards, Vaclav class in_place_factory0: public in_place_factory_base { public: in_place_factory0() {} template<class T> void apply(void* address) const { new(address) T(); } }; in_place_factory0 in_place() { return in_place_factory0(); }

1 0

Monads in boost.mpl?
by Korcan Hussein 10 Mar '06

10 Mar '06

I'm just wondering if there is any specific reason why monadic programming hasn't been explored in pre-processor/template metaprogramming other than (probably) the most obvious reasons. I can see a state monad being probably quite useful. I'm just refreshing my memory on monadic programming in Haskell just to see if this is applicable to problem I'm currently having.

1 0

[Boost.Test] range support
by jth01＠arcor.de 10 Mar '06

10 Mar '06

Hello, what do people think about Boost.Range support? Instead of writing my_range_type r = <complex expression>; std::vector< int > v = list_of(1)(2)(3); BOOST_CHECK_EQUAL_COLLECTIONS(r.begin(), r.end(), v.begin(), v.end()); one could write more conveniently the single line BOOST_CHECK_EQUAL_RANGES( <complex expression>, list_of(1)(2)(3) ); It's also more readable and would so improve the documentational aspect of tests. (Is it ok to post short diffs as this?) Index: boost/test/test_tools.hpp =================================================================== RCS file: /cvsroot/boost/boost/boost/test/test_tools.hpp,v retrieving revision 1.59 diff -u -r1.59 test_tools.hpp --- boost/test/test_tools.hpp 3 Mar 2006 17:39:46 -0000 1.59 +++ boost/test/test_tools.hpp 6 Mar 2006 19:50:52 -0000 @@ -33,6 +33,7 @@ #include <boost/preprocessor/repetition/repeat.hpp> #include <boost/preprocessor/punctuation/comma_if.hpp> #include <boost/preprocessor/arithmetic/add.hpp> +#include <boost/range/functions.hpp> #include <boost/limits.hpp> #include <boost/type_traits/is_array.hpp> @@ -222,6 +223,22 @@ //____________________________________________________________________________// +#define BOOST_EQUAL_RANGES_IMPL( L, R, TL ) \ + BOOST_TEST_TOOL_IMPL( check_impl, ::boost::test_tools::tt_detail::equal_range_impl( \ + (L), (R) ), "", TL, CHECK_EQUAL_COLL ), \ + 2, \ + BOOST_STRINGIZE( L ), BOOST_STRINGIZE( R ) ) \ +/**/ + +#define BOOST_WARN_EQUAL_RANGES( L, R ) \ + BOOST_EQUAL_RANGES_IMPL( L, R, WARN ) +#define BOOST_CHECK_EQUAL_RANGES( L, R ) \ + BOOST_EQUAL_RANGES_IMPL( L, R, CHECK ) +#define BOOST_REQUIRE_EQUAL_RANGES( L, R ) \ + BOOST_EQUAL_RANGES_IMPL( L, R, REQUIRE ) + +//____________________________________________________________________________// + #define BOOST_BITWISE_EQUAL_IMPL( L, R, TL ) \ BOOST_TEST_TOOL_IMPL( check_impl, \ ::boost::test_tools::tt_detail::bitwise_equal_impl( (L), (R) ), \ @@ -547,6 +564,16 @@ //____________________________________________________________________________// +template <typename Left, typename Right> +inline predicate_result +equal_range_impl( Left const& left, Right const& right ) +{ + return equal_coll_impl( boost::begin(left), boost::end(left), + boost::begin(right), boost::end(right) ); +} + +//____________________________________________________________________________// + template <class Left, class Right> inline predicate_result bitwise_equal_impl( Left const& left, Right const& right )

2 2