Boost Users :
Subject: Re: [Boost-users] [Thread] Thread library in boost 1.62 broken with g++ 4.6.3 and -std=c++0x
From: Emil Dotchevski (emildotchevski_at_[hidden])
Date: 2016-10-03 01:27:44
There is an open bug for this, https://svn.boost.org/trac/boost/ticket/11874.
I don't have that version of GCC installed, I'll see if I can install it
tomorrow. The fix suggested by apolukhin is worth a try.
On Sat, Oct 1, 2016 at 7:28 AM, Edward Diener <eldiener_at_[hidden]>
> On 9/29/2016 2:07 AM, Andreas Wehrmann wrote:
>> Hello folks,
>> I've just wanted to try out the new release and noticed that the Thread
>> library was not building.
>> Assuming this code:
>> #include <boost/thread.hpp>
>> int main( void )
>> return 0;
>> When compiled with -std=c++0x the compilation fails with:
>> In file included from
>> from ./boost/boost_1_62_0/boost/exception_ptr.hpp:9,
>> from ./boost/boost_1_62_0/boost/thread/future.hpp:25,
>> from ./boost/boost_1_62_0/boost/thread.hpp:24,
>> from test.cpp:2:
>> ./boost/boost_1_62_0/boost/exception/info.hpp: In instantiation of
>> âboost::error_info<boost::tag_original_exception_type, const
>> instantiated from here
>> ./boost/boost_1_62_0/boost/exception/info.hpp:66:5: error: âclass
>> boost::error_info<boost::tag_original_exception_type, const
>> std::type_info*>â has no member named âvalue_â
>> ./boost/boost_1_62_0/boost/container/detail/pair.hpp:85:30: warning:
>> âboost::container::piecewise_constructâ defined but not used
>> ./boost/boost_1_62_0/boost/system/error_code.hpp:221:36: warning:
>> âboost::system::posix_categoryâ defined but not used [-Wunused-variable]
>> ./boost/boost_1_62_0/boost/system/error_code.hpp:222:36: warning:
>> âboost::system::errno_ecatâ defined but not used [-Wunused-variable]
>> ./boost/boost_1_62_0/boost/system/error_code.hpp:223:36: warning:
>> âboost::system::native_ecatâ defined but not used [-Wunused-variable]
>> Compilation works if I don't use -std=c++0x.
>> The exact compiler version is "g++ (Ubuntu/Linaro 4.6.3-1ubuntu5) 4.6.3".
>> I'm aware that g++ 4.6.3 is not among the compilers tested for C++11,
>> but it is for pre-Cpp11;
>> I'm wondering if there is a reason for that which I might not be aware
>> of (any known problems?).
> This looks like an rvalue references bug in gcc-4.6.3 since error_info
> definitely has a member named 'value_'.
Boost-users list run by williamkempf at hotmail.com, kalb at libertysoft.com, bjorn.karlsson at readsoft.com, gregod at cs.rpi.edu, wekempf at cox.net