August 2012 - Boost - lists.boost.org

[rational] Over/underflow checked update ready for comments
by Dan Searles 27 Aug '12

27 Aug '12

>> Quick overview of the update: the original usage and >> code of the template class remains mostly as before, ie >> boost::rational<int> RationalVar; should work as >> in the original version. Adding 'true' as a second >> template parameter, however, allows an exception to >> be raised if an operation with a class variable results >> in a value that can't accurately be represented >I haven't looked at anything else, but I dislike using true/false as the template argument. I'd prefer meaningful names. _____ >Rob Stewart robert.stewart(a)sig.com >Software Engineer using std::disclaimer; >Dev Tools & Components >Susquehanna International Group, LLP http://www.sig.com Replaced the bool template parameter with an enum: enum RationalChecktype {RationalNoChecking, RationalCheckforOverflow};

1 0

[contract] "Duplicated globals" problem
by Andrzej Krzemienski 26 Aug '12

26 Aug '12

Hi Lorenzo, I think that installing broken contract handlers will not work correctly when a program is linking with DLLs on Windows. I do not know how dynamic linking works on Linux. Here is a sketch of a sample program that fails to work as expected. The program consists of two projects: one builds a DLL, the other builds an executable that links with the DLL: DLL: ==================================== # include <contract.hpp> # include "testdll.h" CONTRACT_FUNCTION ( int (give1) () postcondition(auto result = return, result == 1) ) { return 2; // obviously wrong } DLL_MODE void test_contracts() { int i = give1(); } ==================================== EXE: ==================================== # include <contract.hpp> # include "../testdll/testdll.h" void PostconditionHandler( contract::from const& context ) { std::cerr << "ERROR!!!" << std::endl; } int main() { contract::set_postcondition_broken(&PostconditionHandler); test_contracts(); } ==================================== I expect that the program would use the broken postcondition handler I installed, but it still uses the default one. The cause is that Contract is using some global object to store the installed handler, but when a program is linked with a DLL both the program and the DLL store their own copy of the global. The global is duplicated. THe one in the EXE registers my callback, but the global in the DLL still keeps the default handler. The only way to solve this issue is to -- at least in the cases when users' programs link with DLLs -- ship Contract with a small DLL that stores the unique copy of the globals. This is how Boost.Log handles this problem. Quoting the documentation ( http://boost-log.sourceforge.net/libs/log/doc/html/log/installation.html) *"If your application consists of more than one module (e.g. an exe and one or several dll's) that use Boost.Log, the library must be built as a shared object. If you have a single executable, you may build the library as a static library." * Regards, &rzej* *

2 1

complex acos and atanh
by Stephen Montgomery-Smith 26 Aug '12

26 Aug '12

Hey people, I was trying out your complex arc-trig functions, to compare them with some that I have written: http://people.freebsd.org/~stephen/catrig.c. I am finding that your acos function is getting some of them wrong. For example, your program evaluates acos(1.00000002785941 + I*5.72464869028403e-200) as 0 - I*0.000236048349018331 whereas it should be 2.42520172707401e-196 - I*0.000236048349018331. Also your complex atanh function does quite badly in some edge cases. I presume it is because Hull, Fairgrieve and Tang did not write a paper on that function. If you email Fairgrieve, he may give you a copy of the paper that was never published. He sent me a copy, but doesn't want it widely distributed. I devised my own algorithm at http://people.freebsd.org/~stephen/catrig.c, which actually turns out to be very similar to the unpublished algorithm of Hull, Fairgrieve and Tang. Anyway, I would be happy to work with whomever is currently working on these algorithms. I think someone must be, because I see that someone fixed a bug in asin since 1.48.0. Best regards, Stephen

2 2

[locale] Custom Gettext paths.
by Jookia 26 Aug '12

26 Aug '12

Hello! I'm extremely happy about Boost.Locale, but I've found a few things lacking. So at first I wrote a some wrappers to get around a few flaws with my usage of Boost.Locale. One of these flaws was how it's hardcoded to use the Gettext directory hierarchy. I'd rather store my stuff in 'lang/en_US.mo' rather than 'lang/en_US/LC_MESSAGES/my_app.mo'. The patch adds a 'path format' feature, which allows you to format the directory structure when finding Gettext catalogs, to achieve the effect above. All you really have to do is run: gen.add_path_format("{1}/{2}.mo"); // Use a smaller hierarchy. to achieve the result that I prefer, or gen.add_path_format("{1}/{2}/{3}/{4}.mo"); // Use a Gettext hierarchy. to achieve the result that Boost.Locale uses right now. Ripped straight from Doxygen comments: {1} A path to search for catalogs in. {2} The locale's name. {3} The locale's category. {4} The Gettext domain. I'm new to submitting patches and things, and I'm not sure how I could make this feature better or even acceptable if it isn't already, but I'm sure it'd help others out there. There are a few drawbacks I've found with my code: - To be backwards compatible I've had to add some hackery to add the Gettext format if the path formats vector is empty. - Without consulting documentation, the format is unreadable. - I haven't tested it on the Win32 backend. I'm 90% sure the code would work, it's only 4 lines of code added and it's identical to the code added in the other backends, but it's still untested code. - It's not tested properly. By properly, I mean the tests past with adding a format, and falling back to the default format, but I haven't tried having a 'compact' format. I'm not sure if this would indicate a fault in the formatting code though, as that'd be the only difference code-wise, unless there's more code that has the Gettext hierarchy hardcoded, that didn't fail when I didn't fix the tests. The patch itself is based off of SVN r80097, and despite the diff being for r80104 I haven't tested it in that.. I did the work internally in Git but I made the patch using SVN, in case that matters. I've attached the patch, but I'm not sure how well that's going to go over considering I'm new to mailing lists, and I'm using.. Thunderbird.. Yeah. In case the attachment fails, it's available at http://sprunge.us/VSiX . I'd also like to know if it's appropriate to put '[patch]' in the title for any future patches, and if it should go after '[locale]' or at the end of the title. Thanks, Jookia.

2 13

boost::filesystem -- using non-POSIX locale crashes on Solaris
by JonathonS 26 Aug '12

26 Aug '12

Hi, I am currently using boost::filesystem::path on Solaris, and I noticed that if I set the locale to anything except "C" or "POSIX" such as "en_US.UTF-8", that it crashes i.e. a run-time error gets thrown. I found a link describing the issue here: https://svn.boost.org/trac/boost/ticket/5928 My question is though, whether it is possible to use boost::filesystem without locales? Meaning, if I know my string is a UTF-8 string, is it possible to avoid the locale processing within the boost filesystem? Thanks, J

3 2

[Review] Boost.Contract Review Starts Today(*)
by Dave Abrahams 26 Aug '12

26 Aug '12

The review period for Lorenzo Caminiti's Boost.Contract library runs from today through August 31st. I'm pleased to serve as review manager. ,----[ Important Notes ] | - Please begin your review's subject line with "[Review:Algorithms]" | - Please do not post your review or discussion as a reply to this | message. | - Please feel free to note any individual issues in the issue | tracker at https://sourceforge.net/apps/trac/contractpp/report/3 | (you'll need to log in at Sourceforge to add a ticket) and make | reference to them from your review. | - Please see below for *further* important notes `---- THE LIBRARY Boost.Contract implements Contract Programming (a.k.a., Design by Contracts) for C++. In addition, the library implements virtual specifiers, concept checking, and named parameters. Documentation: http://contractpp.sourceforge.net Source: http://sourceforge.net/projects/contractpp/files/latest/download All N1962 requirements and Eiffel features are supported by the library, among others: 1) Support for preconditions, postconditions, class invariants, block invariants, and loop variants. 2) Subcontract derived classes (with support for pure virtual functions and multiple inheritance). 3) Access expression old values and function return value in postconditions. 4) Optional compilation and checking of preconditions, postconditions, class invariants, block invariants, and loop variants. 5) Customizable actions on contract assertion failure (terminate by default but it can throw, exit, etc). This library is implemented for C++03 and it does not require C++11. Obviously, if the library is accepted, all macros will be prefixed by BOOST_... and other symbols will go under the boost::... namespace. Potential Interactions with Existing Boost Libraries ==================================================== This submission potentially interacts with the following existing Boost libraries: * parameter - http://boost.org/libs/parameter * static_assert - http://www.boost.org/libs/static_assert * MPL - http://www.boost.org/libs/mpl * Concept - http://www.boost.org/libs/concept_check Reviewers should consider these potential interactions. What to include in Review Comments ================================== - Keep your criticism constructive. Constructive criticism usually comes with specific suggestions for improvement. - If you identify problems along the way, please note if they are minor, serious, or showstoppers. - Please *do* vote on whether the library should be accepted. Everyone who takes the time to look at the library is qualified to vote, and every vote is valuable. - Other questions you might want to answer in your review: * 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? (*) Well, OK, the review period officially started yesterday. Sorry for the delayed announcement -- Dave Abrahams BoostPro Computing Software Development Training http://www.boostpro.com Clang/LLVM/EDG Compilers C++ Boost

2 2

[iterator] transform_iterator and move-only types
by Nathan Ridge 25 Aug '12

25 Aug '12

Hi, The following code fails to compile for me: #include <memory> #include <boost/iterator/transform_iterator.hpp> std::unique_ptr<int> foo(int i) { return std::unique_ptr<int>(new int(i)); } int main() { int a[5] = {1, 2, 3, 4, 5}; auto it = boost::make_transform_iterator(a, foo); } The error message I get with GCC 4.8 is pasted below. I think this should work because returning a move-only type from a function by value is a perfectly valid thing to do. Do you agree? Should I file a bug? Thanks, Nate In file included from boost/iterator/iterator_categories.hpp:20:0, from boost/iterator/iterator_adaptor.hpp:14, from boost/iterator/transform_iterator.hpp:12, from test.cpp:2: boost/type_traits/is_convertible.hpp: In instantiation of 'const bool boost::detail::is_convertible_basic_impl<std::unique_ptr<int>&, std::unique_ptr<int> >::value': boost/type_traits/is_convertible.hpp:296:5: required from 'const bool boost::detail::is_convertible_impl<std::unique_ptr<int>, std::unique_ptr<int> >::value' boost/type_traits/is_convertible.hpp:419:1: required from 'struct boost::is_convertible<std::unique_ptr<int>, std::unique_ptr<int> >' boost/mpl/aux_/nested_type_wknd.hpp:26:31: required from 'struct boost::mpl::aux::nested_type_wknd<boost::is_convertible<std::unique_ptr<int>, std::unique_ptr<int> > >' boost/mpl/aux_/preprocessed/gcc/and.hpp:23:8: required from 'struct boost::mpl::aux::and_impl<true, boost::is_convertible<std::unique_ptr<int>, std::unique_ptr<int> >, mpl_::bool_<true>, mpl_::bool_<true>, mpl_::bool_<true> >' boost/mpl/aux_/preprocessed/gcc/and.hpp:48:8: required from 'struct boost::mpl::and_<boost::is_convertible<boost::random_access_traversal_tag, boost::single_pass_traversal_tag>, boost::is_convertible<std::unique_ptr<int>, std::unique_ptr<int> >, mpl_::bool_<true>, mpl_::bool_<true>, mpl_::bool_<true> >' boost/mpl/if.hpp:67:11: [ skipping 6 instantiation contexts, use -ftemplate-backtrace-limit=0 to disable ] boost/iterator/detail/facade_iterator_category.hpp:187:8: required from 'struct boost::detail::facade_iterator_category<boost::random_access_traversal_tag, std::unique_ptr<int>, std::unique_ptr<int> >' boost/iterator/iterator_facade.hpp:104:17: required from 'struct boost::detail::iterator_facade_types<std::unique_ptr<int>, boost::random_access_traversal_tag, std::unique_ptr<int>, long int>' boost/iterator/iterator_facade.hpp:621:9: required from 'class boost::iterator_facade<boost::transform_iterator<std::unique_ptr<int> (*)(int), int*, boost::use_default, boost::use_default>, std::unique_ptr<int>, boost::random_access_traversal_tag, std::unique_ptr<int>, long int>' boost/iterator/iterator_adaptor.hpp:261:9: required from 'class boost::iterator_adaptor<boost::transform_iterator<std::unique_ptr<int> (*)(int), int*, boost::use_default, boost::use_default>, int*, std::unique_ptr<int>, boost::use_default, std::unique_ptr<int>, boost::use_default>' boost/iterator/transform_iterator.hpp:74:9: required from 'class boost::transform_iterator<std::unique_ptr<int> (*)(int), int*, boost::use_default, boost::use_default>' test.cpp:12:52: required from here boost/type_traits/is_convertible.hpp:136:23: error: use of deleted function 'std::unique_ptr<_Tp, _Dp>::unique_ptr(const std::unique_ptr<_Tp, _Dp>&) [with _Tp = int; _Dp = std::default_delete<int>]' static bool const value = sizeof( boost::detail::checker<To>::_m_check(_m_from, 0) ) ^ In file included from c++/4.8.0/memory:81:0, from test.cpp:1: c++/4.8.0/bits/unique_ptr.h:259:7: error: declared here unique_ptr(const unique_ptr&) = delete; ^ In file included from boost/iterator/iterator_categories.hpp:20:0, from boost/iterator/iterator_adaptor.hpp:14, from boost/iterator/transform_iterator.hpp:12, from test.cpp:2: boost/type_traits/is_convertible.hpp:129:41: error: initializing argument 1 of 'static boost::type_traits::yes_type boost::detail::checker<T>::_m_check(T, int) [with T = std::unique_ptr<int>; boost::type_traits::yes_type = char]' static boost::type_traits::yes_type _m_check(T, int); ^

2 2

[tools/regression + filesystem] compiler error
by Jeffrey Lee Hellrung, Jr. 25 Aug '12

25 Aug '12

I was trying to build the boost tools (v1.51.0) and ran into the error pasted below. Should this be concerning? - Jeff Performing configuration checks - has_icu builds : no ...patience... ...patience... ...found 2720 targets... ...updating 3 targets... compile-c-c++ ..\build\boost\regression\msvc-9.0\release\link-static\threading-multi\library_status.obj library_status.cpp regression\src\library_status.cpp(633) : error C2440: 'initializing' : cannot convert from 'const boost::filesystem::path::string_type' to 'std::basic_string<_Elem,_Traits,_Ax>' with [ _Elem=char, _Traits=std::char_traits<char>, _Ax=std::allocator<char> ] No constructor could take the source type, or constructor overload resolution was ambiguous regression\src\library_status.cpp(761) : error C2664: 'std::basic_string<_Elem,_Traits,_Ax> &std::basic_string<_Elem,_Traits,_Ax>::append(const std::basic_string<_Elem,_Traits,_Ax> &)' : cannot convert parameter 1 from 'const boost::filesystem::path::string_type' to 'const std::basic_string<_Elem,_Traits,_Ax> &' with [ _Elem=char, _Traits=std::char_traits<char>, _Ax=std::allocator<char> ] Reason: cannot convert from 'const boost::filesystem::path::string_type' to 'const std::basic_string<_Elem,_Traits,_Ax>' with [ _Elem=char, _Traits=std::char_traits<char>, _Ax=std::allocator<char> ] No user-defined-conversion operator available that can perform this conversion, or the operator cannot be called ..\boost/test/impl/execution_monitor.ipp(948) : warning C4535: calling _set_se_translator() requires /EHa regression\src\library_status.cpp(899) : error C2679: binary '=' : no operator found which takes a right-hand operand of type 'const boost::filesystem::path::string_type' (or there is no acceptable conversion) c:\Program Files (x86)\Microsoft Visual Studio 9.0\VC\INCLUDE\xstring(914): could be 'std::basic_string<_Elem,_Traits,_Ax> &std::basic_string<_Elem,_Traits,_Ax>::operator =(const std::basic_string<_Elem,_Traits,_Ax> &)' with [ _Elem=char, _Traits=std::char_traits<char>, _Ax=std::allocator<char> ] c:\Program Files (x86)\Microsoft Visual Studio 9.0\VC\INCLUDE\xstring(919): or 'std::basic_string<_Elem,_Traits,_Ax> &std::basic_string<_Elem,_Traits,_Ax>::operator =(const _Elem *)' with [ _Elem=char, _Traits=std::char_traits<char>, _Ax=std::allocator<char> ] c:\Program Files (x86)\Microsoft Visual Studio 9.0\VC\INCLUDE\xstring(924): or 'std::basic_string<_Elem,_Traits,_Ax> &std::basic_string<_Elem,_Traits,_Ax>::operator =(_Elem)' with [ _Elem=char, _Traits=std::char_traits<char>, _Ax=std::allocator<char> ] while trying to match the argument list '(std::string, const boost::filesystem::path::string_type)' call "C:\Program Files (x86)\Microsoft Visual Studio 9.0\VC\vcvarsall.bat" x86 >nul cl /Zm800 -nologo @"..\build\boost\regression\msvc-9.0\release\link-static\threading-multi\library_status.obj.rsp" ...failed compile-c-c++ ..\build\boost\regression\msvc-9.0\release\link-static\threading-multi\library_status.obj... ...skipped <p..\build\boost\regression\msvc-9.0\release\link-static\threading-multi>library_status.exe for lack of <p..\build\boost\regression\msvc-9.0\release\link-static\threading-multi>library_status.obj... ...skipped <p..\dist\bin>library_status.exe for lack of <p..\build\boost\regression\msvc-9.0\release\link-static\threading-multi>library_status.exe... ...failed updating 1 target... ...skipped 2 targets...

1 0

[phoenix] Help needed with a custom terminal
by Andrey Semashev 25 Aug '12

25 Aug '12

Hi, I'm trying to define a custom terminal that acts as a replacement of operator<< in expressions when the right argument of the operator is of my special type. The terminal aggregates its left subexpression which results in reference to a stream, and result type of the terminal is the same reference. In short, it looks like this: namespace my { struct argument { std::string m_value; explicit argument(std::string const& value); }; template< typename LeftT > class output_terminal { public: typedef void _is_my_terminal; //! Self type typedef output_terminal this_type; //! Result type definition template< typename > struct result; template< typename ContextT > struct result< this_type(ContextT) > { typedef typename remove_cv< typename remove_reference< ContextT >::type >::type context_type; typedef typename phoenix::evaluator::impl< typename LeftT::proto_base_expr&, context_type, phoenix::unused >::result_type type; }; template< typename ContextT > struct result< const this_type(ContextT) > { typedef typename remove_cv< typename remove_reference< ContextT >::type >::type context_type; typedef typename phoenix::evaluator::impl< typename LeftT::proto_base_expr const&, context_type, phoenix::unused >::result_type type; }; private: //! Left argument actor LeftT m_left; //! Right argument std::string m_right; public: //! Initializing constructor output_terminal(LeftT const& left, std::string const& right); //! Invokation operator template< typename ContextT > typename result< const this_type(ContextT) >::type operator() (ContextT const& ctx) const { typedef typename result< const this_type(ContextT) >::type result_type; result_type strm = phoenix::eval(m_left, ctx); strm << m_right; return strm; } }; template< typename LeftExprT > inline phoenix::actor< output_terminal< phoenix::actor< LeftExprT > > > operator<< (phoenix::actor< LeftExprT > const& left, argument const& right) { phoenix::actor< output_terminal< phoenix::actor< LeftExprT > > > res = { output_terminal< phoenix::actor< LeftExprT > >(left, right.m_value) }; return res; } typedef phoenix::expression::argument< 1 >::type stream_type; const stream_type stream = {}; } // namespace my int main(int, char*[]) { function< void (std::ostream&) > func; func = my::operator<< (my::stream, my::argument("Hello, world!")); func(std::cout); return 0; } I also have specialized phoenix::result_of::is_nullary, phoenix::is_custom_terminal and phoenix::custom_terminal based on the _is_my_terminal tag. The problem is that the code doesn't compile. The compiler output is quite lengthy and it ends in this: ./boost/fusion/adapted/struct/detail/at_impl.hpp:24:16: error: invalid use of incomplete type ‘struct boost::fusion::extension::access::struct_member<boost::phoenix::vector1<const boost::phoenix::actor<boost::proto::exprns_::expr<boost::proto::tagns_::tag::terminal, boost::proto::argsns_::term<my::output_terminal<boost::phoenix::actor<boost::proto::exprns_::basic_expr<boost::proto::tagns_::tag::terminal, boost::proto::argsns_::term<boost::phoenix::argument<1> >, 0l> > > >, 0l> >*>, 1>’ It is not clear where the problem is but I think for some reason Boost.Phoenix thinks that output_terminal is a nullary terminal, despite the fact that I specialized is_nullary to return false. Is there something wrong with my code or is it a problem with Boost.Phoenix? I have posted the complete code here: http://pastebin.com/Vniq1b5Y and here is the compiler output: http://pastebin.com/8Beg4LiY I'm using release SVN branch and GCC 4.6. Thanks in advance.

1 1

[rational] Over/underflow checked update ready for comments
by Dan Searles 25 Aug '12

25 Aug '12

For those interested in rational, there is an updated version in https://svn.boost.org/svn/boost/sandbox/rational ready for comments. Quick overview of the update: the original usage and code of the template class remains mostly as before, ie boost::rational<int> RationalVar; should work as in the original version. Adding 'true' as a second template parameter, however, allows an exception to be raised if an operation with a class variable results in a value that can't accurately be represented, ie boost::rational<int,true> RationalVarChecked. Checking is only needed on limited precision integer types, and it has been tested with signed char, short, int and long long. Thank you for your constructive comments. Dan Searles

2 1