Broken unary_functions in numeric conversions
SHA-1: 2009f0d4fc49a6db59b950a3140f8d1f58e1a8c6 * Revert "removed deprecated std::unary_function from various conversion functors" This reverts commit 593aafab5a7aba86f14df5f9a14a53fc4353bb00. Without this commit boost\libs\numeric\conversion\include\boost\numeric\conversion\detail\converter.hpp is broken on MSVC C++17. (See https://svn.boost.org/trac10/ticket/12972) This is still fixed in develop, does anyone know why the changes were reverted? Can we get someone to actually recommit a fix? Is numeric conversion abandoned? It doesn't look like it has been touched in over a year. Thanks
On 20 August 2017 at 23:05, Gary Furnish via Boost <boost@lists.boost.org> wrote:
SHA-1: 2009f0d4fc49a6db59b950a3140f8d1f58e1a8c6
* Revert "removed deprecated std::unary_function from various conversion functors"
This reverts commit 593aafab5a7aba86f14df5f9a14a53fc4353bb00.
Without this commit boost\libs\numeric\conversion\include\boost\numeric\ conversion\detail\converter.hpp is broken on MSVC C++17. (See https://svn.boost.org/trac10/ticket/12972) This is still fixed in develop, does anyone know why the changes were reverted? Can we get someone to actually recommit a fix? Is numeric conversion abandoned? It doesn't look like it has been touched in over a year. Thanks
I don't know why it have been reverted, but I wonder: is it supposed to work with C++11 and previous if you don't have these inheritance? Should it be conditional? I suspect it might have been the reason. Joël Lamotte
On 22/08/2017 11:19, Klaim - Joël Lamotte via Boost wrote:
On 20 August 2017 at 23:05, Gary Furnish via Boost <boost@lists.boost.org> wrote:
SHA-1: 2009f0d4fc49a6db59b950a3140f8d1f58e1a8c6
* Revert "removed deprecated std::unary_function from various conversion functors"
This reverts commit 593aafab5a7aba86f14df5f9a14a53fc4353bb00.
Without this commit boost\libs\numeric\conversion\include\boost\numeric\ conversion\detail\converter.hpp is broken on MSVC C++17. (See https://svn.boost.org/trac10/ticket/12972) This is still fixed in develop, does anyone know why the changes were reverted? Can we get someone to actually recommit a fix? Is numeric conversion abandoned? It doesn't look like it has been touched in over a year. Thanks
I don't know why it have been reverted, but I wonder: is it supposed to work with C++11 and previous if you don't have these inheritance? Should it be conditional? I suspect it might have been the reason.
This is hitting me as well, and I'd like it fixed, in fact I think I asked about the revert on GitHub, but haven't got a reply. There was some discussion about it here: https://lists.boost.org/Archives/boost/2017/04/234575.php but apparently no resolution :( --- This email has been checked for viruses by AVG. http://www.avg.com
On 08/22/2017 01:45 PM, John Maddock via Boost wrote:
This is hitting me as well, and I'd like it fixed, in fact I think I asked about the revert on GitHub, but haven't got a reply. There was
On 22 August 2017 at 15:06, Bjorn Reese via Boost <boost@lists.boost.org> wrote:
On 08/22/2017 01:45 PM, John Maddock via Boost wrote:
This is hitting me as well, and I'd like it fixed, in fact I think I asked
about the revert on GitHub, but haven't got a reply. There was
Looks like the commit was doing more than what I did to fix my boost version (just remove the inheritance, no addition). I don't have a way to check right now this but maybe the traits are necessary only under c++11. Joël Lamotte
Is there a problem with just merging dev? That seems to work... On Tue, Aug 22, 2017 at 8:25 AM, Brandon Kohn via Boost <boost@lists.boost.org> wrote:
On 8/20/2017 17:05 PM, Gary Furnish via Boost wrote:
SHA-1: 2009f0d4fc49a6db59b950a3140f8d1f58e1a8c6
Can we get someone to actually recommit a fix?
I'm working on this now. Sorry for the delay in responding.
Cheers,
Brandon
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Actually this isn't true at all: develop gets In file included from ../../../../boost/numeric/conversion/detail/numeric_cast_traits.hpp:12:0, from ../../../../boost/numeric/conversion/numeric_cast_traits.hpp:28, from ../../../../boost/numeric/conversion/cast.hpp:34, from compile_fail/built_in_numeric_cast_traits.cpp:11: ../../../../boost/numeric/conversion/detail/preprocessed/numeric_cast_traits_common.hpp:987:12: error: previous definition of ‘struct boost::numeric::numeric_cast_traits<unsigned int, float>’ struct numeric_cast_traits ^~~~~~~~~~~~~~~~~~~ < ~ unsigned int ~~~~~~~~~~~~ , float ~~~~~~~ > ~ compile_fail/built_in_numeric_cast_traits.cpp:24:8: error: redefinition of ‘struct boost::numeric::numeric_cast_traits<long int, float>’ struct numeric_cast_traits< \ ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ BOOST_PP_SEQ_ELEM( BOOST_PP_TUPLE_ELEM(4,0,state) \ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ , BOOST_PP_TUPLE_ELEM(4,3,state) ) \ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ , BOOST_PP_TUPLE_ELEM(4,2,state)> and a bunch more spam. On Tue, Aug 22, 2017 at 8:32 AM, Gary Furnish <gfurnish@gfurnish.net> wrote:
Is there a problem with just merging dev? That seems to work...
On Tue, Aug 22, 2017 at 8:25 AM, Brandon Kohn via Boost <boost@lists.boost.org> wrote:
On 8/20/2017 17:05 PM, Gary Furnish via Boost wrote:
SHA-1: 2009f0d4fc49a6db59b950a3140f8d1f58e1a8c6
Can we get someone to actually recommit a fix?
I'm working on this now. Sorry for the delay in responding.
Cheers,
Brandon
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
participants (5)
-
Bjorn Reese -
Brandon Kohn -
Gary Furnish -
John Maddock -
Klaim - Joël Lamotte