[Boost-users] [boost][multiprecision] compile error with gcc 4.8.4

Rob Conde

unread,

Aug 15, 2016, 8:59:59 PM8/15/16

to boost...@lists.boost.org

Please see the scenario below...am I doing something wrong or is this a gcc or multiprecision bug?

Thanks,

Rob Conde

given the following code (multiPrecisionTest.cpp):

#include <boost/multiprecision/cpp_int.hpp>

#include <boost/multiprecision/cpp_bin_float.hpp>

typedef boost::multiprecision::int128_t int128_t;

typedef boost::multiprecision::cpp_bin_float_quad float128_t;

int main()

{

int128_t intVal;

float128_t floatVal(intVal.convert_to<float128_t>());

return 0;

}

--------------------------------------------------------------------------

g++ -std=c++0x -c multiPrecisionTest.cpp

I get the following error:

In file included from ./boost/multiprecision/cpp_int.hpp:1936:0,

from multiPrecisionTest.cpp:1:

./boost/multiprecision/cpp_int/misc.hpp: In instantiation of ‘typename boost::enable_if_c<(boost::multiprecision::backends::is_trivial_cpp_int<boost::multiprecision::backends::cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1> >::value && boost::multiprecision::is_signed_number<boost::multiprecision::backends::cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1> >::value)>::type boost::multiprecision::backends::eval_convert_to(R*, const boost::multiprecision::backends::cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1>&) [with R = boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>; unsigned int MinBits1 = 128u; unsigned int MaxBits1 = 128u; boost::multiprecision::cpp_integer_type SignType1 = (boost::multiprecision::cpp_integer_type)1u; boost::multiprecision::cpp_int_check_type Checked1 = (boost::multiprecision::cpp_int_check_type)0u; Allocator1 = void; typename boost::enable_if_c<(boost::multiprecision::backends::is_trivial_cpp_int<boost::multiprecision::backends::cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1> >::value && boost::multiprecision::is_signed_number<boost::multiprecision::backends::cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1> >::value)>::type = void]’:

./boost/multiprecision/number.hpp:585:40: required from ‘void boost::multiprecision::number<Backend, ExpressionTemplates>::convert_to_imp(T*) const [with T = boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>; Backend = boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void>; boost::multiprecision::expression_template_option ExpressionTemplates = (boost::multiprecision::expression_template_option)0u]’

./boost/multiprecision/number.hpp:601:29: required from ‘T boost::multiprecision::number<Backend, ExpressionTemplates>::convert_to() const [with T = boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>; Backend = boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void>; boost::multiprecision::expression_template_option ExpressionTemplates = (boost::multiprecision::expression_template_option)0u]’

multiPrecisionTest.cpp:10:54: required from here

./boost/multiprecision/cpp_int/misc.hpp:533:138: error: no type named ‘type’ in ‘struct boost::common_type<boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>, __int128 unsigned>’

typedef typename common_type<R, typename cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1>::local_limb_type>::type common_type;

^

John Maddock

unread,

Aug 16, 2016, 7:11:29 AM8/16/16

to boost...@lists.boost.org

On 16/08/2016 01:59, Rob Conde wrote:
> #include <boost/multiprecision/cpp_int.hpp>
> #include <boost/multiprecision/cpp_bin_float.hpp>
>
> typedef boost::multiprecision::int128_t int128_t;
> typedef boost::multiprecision::cpp_bin_float_quad float128_t;
>
> int main()
> {
> int128_t intVal;
> float128_t floatVal(intVal.convert_to<float128_t>());
> return 0;
> }

That should work, as should:

float128_t f(intVal);

However, while both of these do work with msvc, they fail with GCC, so
there is a bug (or 2) somewhere. I'll investigate later.

John.
_______________________________________________
Boost-users mailing list
Boost...@lists.boost.org
http://lists.boost.org/mailman/listinfo.cgi/boost-users

Rob Conde

unread,

Aug 18, 2016, 10:06:00 AM8/18/16

to boost...@lists.boost.org

Any luck? It would be great if there was a fix in boost 1.62.

Thanks,

Rob

From: Boost-users <boost-use...@lists.boost.org> on behalf of John Maddock <jz.ma...@googlemail.com>
Sent: Tuesday, August 16, 2016 7:11:07 AM
To: boost...@lists.boost.org
Subject: Re: [Boost-users] [boost][multiprecision] compile error with gcc 4.8.4

John Maddock

unread,

Aug 18, 2016, 12:14:53 PM8/18/16

to boost...@lists.boost.org

On 18/08/2016 15:05, Rob Conde wrote:
>
> Any luck? It would be great if there was a fix in boost 1.62.
>

Fixed in
https://github.com/boostorg/multiprecision/commit/339818c0180d0719852b3e648b8d76f2ba02e32d

If the tests cycle OK, I'll try to get this into 1.62, but it's tight now.

John.

>
> Thanks,
>
> Rob
>
> ------------------------------------------------------------------------
> *From:* Boost-users <boost-use...@lists.boost.org> on behalf of
> John Maddock <jz.ma...@googlemail.com>
> *Sent:* Tuesday, August 16, 2016 7:11:07 AM
> *To:* boost...@lists.boost.org
> *Subject:* Re: [Boost-users] [boost][multiprecision] compile error

Rob Conde

unread,

Aug 18, 2016, 12:17:16 PM8/18/16

to boost...@lists.boost.org

Thanks!

Rob Conde

From: Boost-users <boost-use...@lists.boost.org> on behalf of John Maddock <jz.ma...@googlemail.com>
Sent: Thursday, August 18, 2016 12:14:06 PM
To: boost...@lists.boost.org
Subject: Re: [Boost-users] [boost][multiprecision] compile error with gcc 4.8.4

Rob Conde

unread,

Sep 15, 2016, 12:21:12 PM9/15/16

to boost...@lists.boost.org

I tried with 1.62 beta and ran into the following problems:

Building with std=c++98 for the following case:

--------------------------------------------------------------------------------

#include <boost/multiprecision/cpp_int.hpp>

#include <boost/multiprecision/cpp_bin_float.hpp>

typedef boost::multiprecision::int128_t int128_t;

typedef boost::multiprecision::cpp_bin_float_quad float128_t;

void func()

{

int128_t iVal;

float128_t floatVal(iVal.convert_to<float128_t>());

}

--------------------------------------------------------------------------------

In file included from ./boost/type_traits/detail/common_type_impl.hpp:12:0,

from ./boost/type_traits/common_type.hpp:18,

from ./boost/multiprecision/cpp_int.hpp:22,

from mpTest.cpp:1:

./boost/type_traits/detail/common_arithmetic_type.hpp: In instantiation of ‘class boost::type_traits_detail::common_arithmetic_type<boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>, __int128 unsigned>’:

./boost/type_traits/detail/common_type_impl.hpp:99:35: required from ‘struct boost::type_traits_detail::common_type_impl5<boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>, __int128 unsigned>’

./boost/type_traits/detail/common_type_impl.hpp:90:35: required from ‘struct boost::type_traits_detail::common_type_impl4<boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>, __int128 unsigned>’

./boost/type_traits/detail/common_type_impl.hpp:79:35: required from ‘struct boost::type_traits_detail::common_type_impl3<boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>, __int128 unsigned>’

./boost/type_traits/detail/common_type_impl.hpp:52:35: required from ‘struct boost::type_traits_detail::common_type_class<boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>, __int128 unsigned>’

./boost/type_traits/detail/common_type_impl.hpp:68:35: required from ‘struct boost::type_traits_detail::common_type_impl<boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>, __int128 unsigned>’

./boost/type_traits/common_type.hpp:133:37: required from ‘struct boost::type_traits_detail::common_type_decay_helper<boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>, __int128 unsigned, boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>, __int128 unsigned>’

./boost/type_traits/common_type.hpp:139:37: required from ‘struct boost::common_type<boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>, __int128 unsigned, void, void, void, void, void, void, void>’

./boost/multiprecision/cpp_int/misc.hpp:547:138: required from ‘typename boost::enable_if_c<(boost::multiprecision::backends::is_trivial_cpp_int<boost::multiprecision::backends::cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1> >::value && boost::multiprecision::is_signed_number<boost::multiprecision::backends::cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1> >::value)>::type boost::multiprecision::backends::eval_convert_to(R*, const boost::multiprecision::backends::cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1>&) [with R = boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>; unsigned int MinBits1 = 128u; unsigned int MaxBits1 = 128u; boost::multiprecision::cpp_integer_type SignType1 = (boost::multiprecision::cpp_integer_type)1u; boost::multiprecision::cpp_int_check_type Checked1 = (boost::multiprecision::cpp_int_check_type)0u; Allocator1 = void; typename boost::enable_if_c<(boost::multiprecision::backends::is_trivial_cpp_int<boost::multiprecision::backends::cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1> >::value && boost::multiprecision::is_signed_number<boost::multiprecision::backends::cpp_int_backend<MinBits1, MaxBits1, SignType1, Checked1, Allocator1> >::value)>::type = void]’

./boost/multiprecision/number.hpp:585:40: required from ‘void boost::multiprecision::number<Backend, ExpressionTemplates>::convert_to_imp(T*) const [with T = boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>; Backend = boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void>; boost::multiprecision::expression_template_option ExpressionTemplates = (boost::multiprecision::expression_template_option)0u]’

./boost/multiprecision/number.hpp:601:29: required from ‘T boost::multiprecision::number<Backend, ExpressionTemplates>::convert_to() const [with T = boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>; Backend = boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void>; boost::multiprecision::expression_template_option ExpressionTemplates = (boost::multiprecision::expression_template_option)0u]’

mpTest.cpp:11:52: required from here

./boost/type_traits/detail/common_arithmetic_type.hpp:211:60: error: no match for ternary ‘operator?:’ (operand types are ‘bool’, ‘boost::multiprecision::number<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>, (boost::multiprecision::expression_template_option)0u>’, and ‘__int128 unsigned’)

typedef typename arithmetic_type< sizeof(select( cond()? T(): U() )) >::type type;

Building without a std specified I get the following case:

--------------------------------------------------------------------------------

#include <boost/multiprecision/cpp_int.hpp>

#include <boost/multiprecision/cpp_bin_float.hpp>

typedef boost::multiprecision::int128_t int128_t;

typedef boost::multiprecision::cpp_bin_float_quad float128_t;

void func()

{

float128_t floatVal;

int128_t roundedNanoseconds(floatVal.convert_to<int128_t>());

}

--------------------------------------------------------------------------------

In file included from ./boost/multiprecision/detail/generic_interconvert.hpp:9:0,

from ./boost/multiprecision/number.hpp:22,

from ./boost/multiprecision/cpp_int.hpp:12,

from mpTest.cpp:1:

./boost/multiprecision/detail/default_ops.hpp: In instantiation of ‘void boost::multiprecision::default_ops::eval_convert_to(boost::multiprecision::default_ops::terminal<boost::multiprecision::number<B1, et> >*, const B2&) [with B1 = boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void>; B2 = boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>; boost::multiprecision::expression_template_option et = (boost::multiprecision::expression_template_option)0u]’:

./boost/multiprecision/detail/default_ops.hpp:896:31: required from ‘void boost::multiprecision::default_ops::eval_convert_to(R*, const B&) [with R = boost::multiprecision::number<boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void> >; B = boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>]’

./boost/multiprecision/number.hpp:585:40: required from ‘void boost::multiprecision::number<Backend, ExpressionTemplates>::convert_to_imp(T*) const [with T = boost::multiprecision::number<boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void> >; Backend = boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>; boost::multiprecision::expression_template_option ExpressionTemplates = (boost::multiprecision::expression_template_option)0u]’

./boost/multiprecision/number.hpp:601:29: required from ‘T boost::multiprecision::number<Backend, ExpressionTemplates>::convert_to() const [with T = boost::multiprecision::number<boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void> >; Backend = boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>; boost::multiprecision::expression_template_option ExpressionTemplates = (boost::multiprecision::expression_template_option)0u]’

mpTest.cpp:11:62: required from here

./boost/multiprecision/detail/default_ops.hpp:922:134: error: no matching function for call to ‘generic_interconvert(boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void>&, const boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383>&, boost::multiprecision::number_category<boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void> >, boost::multiprecision::number_category<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383> >)’

boost::multiprecision::detail::generic_interconvert(result->value.backend(), backend, number_category<B1>(), number_category<B2>());

^

./boost/multiprecision/detail/default_ops.hpp:922:134: note: candidates are:

In file included from ./boost/multiprecision/number.hpp:22:0,

from ./boost/multiprecision/cpp_int.hpp:12,

from mpTest.cpp:1:

./boost/multiprecision/detail/generic_interconvert.hpp:30:6: note: template<class To, class From> void boost::multiprecision::detail::generic_interconvert(To&, const From&, const mpl_::int_<1>&, const mpl_::int_<0>&)

void generic_interconvert(To& to, const From& from, const mpl::int_<number_kind_floating_point>& /*to_type*/, const mpl::int_<number_kind_integer>& /*from_type*/)

^

./boost/multiprecision/detail/generic_interconvert.hpp:30:6: note: template argument deduction/substitution failed:

In file included from ./boost/multiprecision/detail/generic_interconvert.hpp:9:0,

from ./boost/multiprecision/number.hpp:22,

from ./boost/multiprecision/cpp_int.hpp:12,

from mpTest.cpp:1:

./boost/multiprecision/detail/default_ops.hpp:922:134: note: cannot convert ‘boost::multiprecision::number_category<boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void> >()’ (type ‘boost::multiprecision::number_category<boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void> >’) to type ‘const mpl_::int_<1>&’

boost::multiprecision::detail::generic_interconvert(result->value.backend(), backend, number_category<B1>(), number_category<B2>());

^

In file included from ./boost/multiprecision/number.hpp:22:0,

from ./boost/multiprecision/cpp_int.hpp:12,

from mpTest.cpp:1:

./boost/multiprecision/detail/generic_interconvert.hpp:78:6: note: template<class To, class From> void boost::multiprecision::detail::generic_interconvert(To&, const From&, const mpl_::int_<0>&, const mpl_::int_<0>&)

void generic_interconvert(To& to, const From& from, const mpl::int_<number_kind_integer>& /*to_type*/, const mpl::int_<number_kind_integer>& /*from_type*/)

^

./boost/multiprecision/detail/generic_interconvert.hpp:78:6: note: template argument deduction/substitution failed:

In file included from ./boost/multiprecision/detail/generic_interconvert.hpp:9:0,

from ./boost/multiprecision/number.hpp:22,

from ./boost/multiprecision/cpp_int.hpp:12,

from mpTest.cpp:1:

./boost/multiprecision/detail/default_ops.hpp:922:134: note: cannot convert ‘boost::multiprecision::number_category<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383> >()’ (type ‘boost::multiprecision::number_category<boost::multiprecision::backends::cpp_bin_float<113u, (boost::multiprecision::backends::digit_base_type)2u, void, short int, -16382, 16383> >’) to type ‘const mpl_::int_<0>&’

boost::multiprecision::detail::generic_interconvert(result->value.backend(), backend, number_category<B1>(), number_category<B2>());

^

In file included from ./boost/multiprecision/number.hpp:22:0,

from ./boost/multiprecision/cpp_int.hpp:12,

from mpTest.cpp:1:

./boost/multiprecision/detail/generic_interconvert.hpp:126:6: note: template<class To, class From> void boost::multiprecision::detail::generic_interconvert(To&, const From&, const mpl_::int_<1>&, const mpl_::int_<1>&)

void generic_interconvert(To& to, const From& from, const mpl::int_<number_kind_floating_point>& /*to_type*/, const mpl::int_<number_kind_floating_point>& /*from_type*/)

^

./boost/multiprecision/detail/generic_interconvert.hpp:126:6: note: template argument deduction/substitution failed:

In file included from ./boost/multiprecision/detail/generic_interconvert.hpp:9:0,

from ./boost/multiprecision/number.hpp:22,

from ./boost/multiprecision/cpp_int.hpp:12,

from mpTest.cpp:1:

./boost/multiprecision/detail/default_ops.hpp:922:134: note: cannot convert ‘boost::multiprecision::number_category<boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void> >()’ (type ‘boost::multiprecision::number_category<boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void> >’) to type ‘const mpl_::int_<1>&’

boost::multiprecision::detail::generic_interconvert(result->value.backend(), backend, number_category<B1>(), number_category<B2>());

^

In file included from ./boost/multiprecision/number.hpp:22:0,

from ./boost/multiprecision/cpp_int.hpp:12,

from mpTest.cpp:1:

./boost/multiprecision/detail/generic_interconvert.hpp:209:6: note: template<class To, class From> void boost::multiprecision::detail::generic_interconvert(To&, const From&, const mpl_::int_<2>&, const mpl_::int_<2>&)

void generic_interconvert(To& to, const From& from, const mpl::int_<number_kind_rational>& /*to_type*/, const mpl::int_<number_kind_rational>& /*from_type*/)

^

./boost/multiprecision/detail/generic_interconvert.hpp:209:6: note: template argument deduction/substitution failed:

In file included from ./boost/multiprecision/detail/generic_interconvert.hpp:9:0,

from ./boost/multiprecision/number.hpp:22,

from ./boost/multiprecision/cpp_int.hpp:12,

from mpTest.cpp:1:

./boost/multiprecision/detail/default_ops.hpp:922:134: note: cannot convert ‘boost::multiprecision::number_category<boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void> >()’ (type ‘boost::multiprecision::number_category<boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void> >’) to type ‘const mpl_::int_<2>&’

boost::multiprecision::detail::generic_interconvert(result->value.backend(), backend, number_category<B1>(), number_category<B2>());

^

In file included from ./boost/multiprecision/number.hpp:22:0,

from ./boost/multiprecision/cpp_int.hpp:12,

from mpTest.cpp:1:

./boost/multiprecision/detail/generic_interconvert.hpp:220:6: note: template<class To, class From> void boost::multiprecision::detail::generic_interconvert(To&, const From&, const mpl_::int_<2>&, const mpl_::int_<0>&)

void generic_interconvert(To& to, const From& from, const mpl::int_<number_kind_rational>& /*to_type*/, const mpl::int_<number_kind_integer>& /*from_type*/)

^

./boost/multiprecision/detail/generic_interconvert.hpp:220:6: note: template argument deduction/substitution failed:

In file included from ./boost/multiprecision/detail/generic_interconvert.hpp:9:0,

from ./boost/multiprecision/number.hpp:22,

from ./boost/multiprecision/cpp_int.hpp:12,

from mpTest.cpp:1:

./boost/multiprecision/detail/default_ops.hpp:922:134: note: cannot convert ‘boost::multiprecision::number_category<boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void> >()’ (type ‘boost::multiprecision::number_category<boost::multiprecision::backends::cpp_int_backend<128u, 128u, (boost::multiprecision::cpp_integer_type)1u, (boost::multiprecision::cpp_int_check_type)0u, void> >’) to type ‘const mpl_::int_<2>&’

boost::multiprecision::detail::generic_interconvert(result->value.backend(), backend, number_category<B1>(), number_category<B2>());

^

Rob

From: Boost-users <boost-use...@lists.boost.org> on behalf of John Maddock <jz.ma...@googlemail.com>
Sent: Thursday, August 18, 2016 12:14:06 PM
To: boost...@lists.boost.org

Subject: Re: [Boost-users] [boost][multiprecision] compile error with gcc 4.8.4

John Maddock

unread,

Sep 16, 2016, 7:55:25 AM9/16/16

to boost...@lists.boost.org

On 15/09/2016 17:20, Rob Conde wrote:
>
> I tried with 1.62 beta and ran into the following problems:
>

I think this fixes these:
https://github.com/boostorg/multiprecision/commit/b2108867c7415cc1ed28244843555289fdb8d356

Or at least it does for me, can you verify with your setup as well please?

Whether there is time for this to make 1.62 is another matter....

John.

> ------------------------------------------------------------------------
> *From:* Boost-users <boost-use...@lists.boost.org> on behalf of
> John Maddock <jz.ma...@googlemail.com>

> *Sent:* Thursday, August 18, 2016 12:14:06 PM

Rob Conde

unread,

Sep 16, 2016, 2:51:32 PM9/16/16

to boost...@lists.boost.org

The first case (int -> float) was fixed, but the second (float->int) wasn't.

Rob

From: Boost-users <boost-use...@lists.boost.org> on behalf of John Maddock <jz.ma...@googlemail.com>
Sent: Friday, September 16, 2016 7:54:26 AM
To: boost...@lists.boost.org
Subject: Re: [Boost-users] [boost][multiprecision] compile error with gcc 4.8.4

John Maddock

unread,

Sep 17, 2016, 7:01:14 AM9/17/16

to boost...@lists.boost.org

On 16/09/2016 19:50, Rob Conde wrote:
>
> The first case (int -> float) was fixed, but the second (float->int)
> wasn't.
>

Oh :(

I fear this may have to stay broken for this release - what I need to do
is devise a test case that tests conversion of everything-to-everything
- and that's going to be a lot of permutations!

John.

>
> Rob
>
> ------------------------------------------------------------------------
> *From:* Boost-users <boost-use...@lists.boost.org> on behalf of
> John Maddock <jz.ma...@googlemail.com>

> *Sent:* Friday, September 16, 2016 7:54:26 AM

John Maddock

unread,

Sep 17, 2016, 7:49:46 AM9/17/16

to boost...@lists.boost.org

On 17/09/2016 11:59, John Maddock wrote:
>
>
> On 16/09/2016 19:50, Rob Conde wrote:
>>
>> The first case (int -> float) was fixed, but the second (float->int)
>> wasn't.
>>
>
> Oh :(
>
> I fear this may have to stay broken for this release - what I need to
> do is devise a test case that tests conversion of
> everything-to-everything - and that's going to be a lot of permutations!

On second thoughts, the second case (float -> int) is a conversion we
have never supported (unsafe/lossy) - which is to say this is by
design. That said, we have always supported conversion to native
(built-in) integer types, so we're inconsistent there.

John.

Rob Conde

unread,

Sep 17, 2016, 9:33:56 AM9/17/16

to boost...@lists.boost.org

On 16/09/2016 19:50, Rob Conde wrote:

The first case (int -> float) was fixed, but the second (float->int) wasn't.

Oh :(

I fear this may have to stay broken for this release - what I need to do is devise a test case that tests conversion of everything-to-everything - and that's going to be a lot of permutations!

On second thoughts, the second case (float -> int) is a conversion we have never supported (unsafe/lossy) - which is to say this is by design. That said, we have always supported conversion to native (built-in) integer types, so we're inconsistent there.

John.

Can you clarify your approach going forward? When this compiles (other versions of gcc, visual studio) it works as intended. It's not a huge deal for me whether this is fixed for 1.62, but I definitely don't want it to be expressly forbidden going forward.

Rob

John Maddock

unread,

Sep 17, 2016, 12:58:44 PM9/17/16

to boost...@lists.boost.org

> Can you clarify your approach going forward? When this compiles (other
> versions of gcc, visual studio) it works as intended. It's not a huge
> deal for me whether this is fixed for 1.62, but I definitely don't
> want it to be expressly forbidden going forward.
>

The float->int conversion fails for me with as expected with msvc and
current gcc versions. Of course the error message should be a lot better.

The ideal solution would be to overload iround/itrunc for multiprecision
integer target types, but of course that's not possible. I guess an
explicit cast is just about acceptable, but let me think about this.

Rob Conde

unread,

Sep 17, 2016, 1:22:37 PM9/17/16

to boost...@lists.boost.org

The float->int conversion fails for me with as expected with msvc and

current gcc versions. Of course the error message should be a lot better.

When you say "fail" do you mean at compile time or runtime?

Rob

Rob Conde

unread,

Sep 17, 2016, 1:36:10 PM9/17/16

to boost...@lists.boost.org

The following works with me in visual studio with toolset v120 and v140.

#include <boost/multiprecision/cpp_int.hpp>

#include <boost/multiprecision/cpp_bin_float.hpp>

#include <iostream>

typedef boost::multiprecision::int128_t int128_t;

typedef boost::multiprecision::cpp_bin_float_quad float128_t;

void main()

{

float128_t floatVal("6.0");

int128_t iVal(floatVal.convert_to<int128_t>());

std::cout << iVal << std::endl;

}

If I change it to have a fractional component, then indeed I get an exception...maybe that's what you mean. I personally only need the above behavior (though keeping the c++ truncation semantics might make sense as a long term goal...but i digress). I always floor/ceil because call convert_to<int128_t>().

Using boost 1.61 and gcc version 4.4.7, the above also compiles and works.

Rob

John Maddock

unread,

Sep 17, 2016, 2:43:46 PM9/17/16

to boost...@lists.boost.org

OK, I see that compiles with 1.61, it does not with current master
though (what will be 1.62).

In 1.61 the code performs a lexical_cast for the conversion - that works
more or less by accident for small integers, but for large ones it fails
at runtime (as an exception) as the numbers get printed in scientific
format during the lexical_cast. Conversion to long long works fine
though, and probably in all the cases that your conversion to int128_t does.

As I said, I need to figure out some more consistent behaviour for
future releases.

Rob Conde

unread,

Sep 17, 2016, 3:58:06 PM9/17/16

to boost...@lists.boost.org

Understood...thanks for your effort.

Rob Conde

Sent from my iPhone

Rob Conde

unread,

Sep 23, 2016, 2:14:15 PM9/23/16

to boost...@lists.boost.org

John,

In the meantime, consider the following workaround:

std::ostringstream o;

o << std::fixed << std::setprecision(1) << tmp;

int128_t roundedNanoseconds(o.str().substr(0,o.tellp() - std::streampos(2)));

Does that seem reasonable, or do you see a much cleaner or efficient approach?

Would be nicer to use setprecision(0), but:

https://svn.boost.org/trac/boost/ticket/10103

Thanks for your input,

Rob Conde

From: Rob Conde
Sent: Saturday, September 17, 2016 3:57:39 PM
To: boost...@lists.boost.org
Subject: Re: [Boost-users] [boost][multiprecision] compile error with gcc 4.8.4

John Maddock

unread,

Sep 24, 2016, 11:32:23 AM9/24/16

to boost...@lists.boost.org

On 23/09/2016 19:13, Rob Conde wrote:
>
> John,
>
> In the meantime, consider the following workaround:
>
>
> std::ostringstream o;
>
> o << std::fixed << std::setprecision(1) << tmp;
>
> int128_t roundedNanoseconds(o.str().substr(0,o.tellp() -
> std::streampos(2)));
>
>
> Does that seem reasonable, or do you see a much cleaner or
> efficient approach?
>

Nod. As a temporary workaround that would work. Longer term a more
efficient solution is to use ldexp/frexp to pull off N bits at a time.

>
> Would be nicer to use setprecision(0), but:
>
>
> https://svn.boost.org/trac/boost/ticket/10103
>

Yes, I know. That's another one I need to work on / think about.

Best, John.

>
> Thanks for your input,
>
> Rob Conde
>

> ------------------------------------------------------------------------
> *From:* Rob Conde
> *Sent:* Saturday, September 17, 2016 3:57:39 PM
> *To:* boost...@lists.boost.org
> *Subject:* Re: [Boost-users] [boost][multiprecision] compile error

> with gcc 4.8.4
> Understood...thanks for your effort.
>
> Rob Conde
>
> Sent from my iPhone
>
> On Sep 17, 2016, at 2:43 PM, John Maddock <jz.ma...@googlemail.com

>> Boost...@lists.boost.org <mailto:Boost...@lists.boost.org>
>> http://lists.boost.org/mailman/listinfo.cgi/boost-users

Reply all

Reply to author

Forward