* Should reinterpret_cast be constexpr?
@ 2020-12-28 19:51 Josh Marshall
2020-12-28 20:33 ` Jonathan Wakely
0 siblings, 1 reply; 6+ messages in thread
From: Josh Marshall @ 2020-12-28 19:51 UTC (permalink / raw)
To: libstdc++
Hello all,
From the test error `libstdc++-v3/include/bits/stl_vector.h:1837: error:
'reinterpret_cast' is not a constant expression` and the documentation at
https://en.cppreference.com/w/cpp/language/reinterpret_cast , is
`reinterpret_cast()` not constexpr in error?
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Should reinterpret_cast be constexpr?
2020-12-28 19:51 Should reinterpret_cast be constexpr? Josh Marshall
@ 2020-12-28 20:33 ` Jonathan Wakely
2020-12-28 20:47 ` Jonathan Wakely
0 siblings, 1 reply; 6+ messages in thread
From: Jonathan Wakely @ 2020-12-28 20:33 UTC (permalink / raw)
To: Josh Marshall; +Cc: libstdc++
On Mon, 28 Dec 2020 at 19:51, Josh Marshall via Libstdc++
<libstdc++@gcc.gnu.org> wrote:
>
> Hello all,
>
> From the test error `libstdc++-v3/include/bits/stl_vector.h:1837: error:
> 'reinterpret_cast' is not a constant expression` and the documentation at
> https://en.cppreference.com/w/cpp/language/reinterpret_cast , is
> `reinterpret_cast()` not constexpr in error?
No, it's not an error. reinterpret_cast is not allowed in constant expressions.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Should reinterpret_cast be constexpr?
2020-12-28 20:33 ` Jonathan Wakely
@ 2020-12-28 20:47 ` Jonathan Wakely
2020-12-28 20:49 ` Josh Marshall
0 siblings, 1 reply; 6+ messages in thread
From: Jonathan Wakely @ 2020-12-28 20:47 UTC (permalink / raw)
To: Josh Marshall; +Cc: libstdc++
On Mon, 28 Dec 2020 at 20:33, Jonathan Wakely <jwakely.gcc@gmail.com> wrote:
>
> On Mon, 28 Dec 2020 at 19:51, Josh Marshall via Libstdc++
> <libstdc++@gcc.gnu.org> wrote:
> >
> > Hello all,
> >
> > From the test error `libstdc++-v3/include/bits/stl_vector.h:1837: error:
> > 'reinterpret_cast' is not a constant expression` and the documentation at
> > https://en.cppreference.com/w/cpp/language/reinterpret_cast , is
> > `reinterpret_cast()` not constexpr in error?
>
> No, it's not an error. reinterpret_cast is not allowed in constant expressions.
This patch replaces the use of reinterpret_cast with an alternative
using C++20 features:
diff --git a/libstdc++-v3/include/bits/vector.tcc
b/libstdc++-v3/include/bits/vector.tcc
index 27e63388feb..93b2ccc02e3 100644
--- a/libstdc++-v3/include/bits/vector.tcc
+++ b/libstdc++-v3/include/bits/vector.tcc
@@ -516,9 +516,13 @@ _GLIBCXX_BEGIN_NAMESPACE_CONTAINER
{
#if __cplusplus < 201103L
value_type __x_copy = __x;
-#else
+#elif ! __cpp_lib_make_obj_using_allocator
_Temporary_value __tmp(this, __x);
value_type& __x_copy = __tmp._M_val();
+#else
+ value_type __x_copy = std::make_from_tuple<value_type>(
+ std::uses_allocator_construction_args<value_type>(
+ _M_get_Tp_allocator(), __x));
#endif
const size_type __elems_after = end() - __position;
pointer __old_finish(this->_M_impl._M_finish);
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Should reinterpret_cast be constexpr?
2020-12-28 20:47 ` Jonathan Wakely
@ 2020-12-28 20:49 ` Josh Marshall
2020-12-28 21:24 ` Jonathan Wakely
0 siblings, 1 reply; 6+ messages in thread
From: Josh Marshall @ 2020-12-28 20:49 UTC (permalink / raw)
To: Jonathan Wakely; +Cc: libstdc++
Is this in master? If so, I'll just merge it in.
On Mon, Dec 28, 2020 at 3:47 PM Jonathan Wakely <jwakely.gcc@gmail.com>
wrote:
> On Mon, 28 Dec 2020 at 20:33, Jonathan Wakely <jwakely.gcc@gmail.com>
> wrote:
> >
> > On Mon, 28 Dec 2020 at 19:51, Josh Marshall via Libstdc++
> > <libstdc++@gcc.gnu.org> wrote:
> > >
> > > Hello all,
> > >
> > > From the test error `libstdc++-v3/include/bits/stl_vector.h:1837:
> error:
> > > 'reinterpret_cast' is not a constant expression` and the documentation
> at
> > > https://en.cppreference.com/w/cpp/language/reinterpret_cast , is
> > > `reinterpret_cast()` not constexpr in error?
> >
> > No, it's not an error. reinterpret_cast is not allowed in constant
> expressions.
>
> This patch replaces the use of reinterpret_cast with an alternative
> using C++20 features:
>
> diff --git a/libstdc++-v3/include/bits/vector.tcc
> b/libstdc++-v3/include/bits/vector.tcc
> index 27e63388feb..93b2ccc02e3 100644
> --- a/libstdc++-v3/include/bits/vector.tcc
> +++ b/libstdc++-v3/include/bits/vector.tcc
> @@ -516,9 +516,13 @@ _GLIBCXX_BEGIN_NAMESPACE_CONTAINER
> {
> #if __cplusplus < 201103L
> value_type __x_copy = __x;
> -#else
> +#elif ! __cpp_lib_make_obj_using_allocator
> _Temporary_value __tmp(this, __x);
> value_type& __x_copy = __tmp._M_val();
> +#else
> + value_type __x_copy = std::make_from_tuple<value_type>(
> + std::uses_allocator_construction_args<value_type>(
> + _M_get_Tp_allocator(), __x));
> #endif
> const size_type __elems_after = end() - __position;
> pointer __old_finish(this->_M_impl._M_finish);
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Should reinterpret_cast be constexpr?
2020-12-28 20:49 ` Josh Marshall
@ 2020-12-28 21:24 ` Jonathan Wakely
[not found] ` <CAFkJGRem_HFTtj93pyWVpvMtjSDLBU8u9nUVXshOTovsEjDaHw@mail.gmail.com>
0 siblings, 1 reply; 6+ messages in thread
From: Jonathan Wakely @ 2020-12-28 21:24 UTC (permalink / raw)
To: Josh Marshall; +Cc: libstdc++
On Mon, 28 Dec 2020 at 20:50, Josh Marshall
<joshua.r.marshall.1991@gmail.com> wrote:
>
> Is this in master? If so, I'll just merge it in.
No, I just made a local change and haven't even committed it locally.
>
> On Mon, Dec 28, 2020 at 3:47 PM Jonathan Wakely <jwakely.gcc@gmail.com> wrote:
>>
>> On Mon, 28 Dec 2020 at 20:33, Jonathan Wakely <jwakely.gcc@gmail.com> wrote:
>> >
>> > On Mon, 28 Dec 2020 at 19:51, Josh Marshall via Libstdc++
>> > <libstdc++@gcc.gnu.org> wrote:
>> > >
>> > > Hello all,
>> > >
>> > > From the test error `libstdc++-v3/include/bits/stl_vector.h:1837: error:
>> > > 'reinterpret_cast' is not a constant expression` and the documentation at
>> > > https://en.cppreference.com/w/cpp/language/reinterpret_cast , is
>> > > `reinterpret_cast()` not constexpr in error?
>> >
>> > No, it's not an error. reinterpret_cast is not allowed in constant expressions.
>>
>> This patch replaces the use of reinterpret_cast with an alternative
>> using C++20 features:
>>
>> diff --git a/libstdc++-v3/include/bits/vector.tcc
>> b/libstdc++-v3/include/bits/vector.tcc
>> index 27e63388feb..93b2ccc02e3 100644
>> --- a/libstdc++-v3/include/bits/vector.tcc
>> +++ b/libstdc++-v3/include/bits/vector.tcc
>> @@ -516,9 +516,13 @@ _GLIBCXX_BEGIN_NAMESPACE_CONTAINER
>> {
>> #if __cplusplus < 201103L
>> value_type __x_copy = __x;
>> -#else
>> +#elif ! __cpp_lib_make_obj_using_allocator
>> _Temporary_value __tmp(this, __x);
>> value_type& __x_copy = __tmp._M_val();
>> +#else
>> + value_type __x_copy = std::make_from_tuple<value_type>(
>> + std::uses_allocator_construction_args<value_type>(
>> + _M_get_Tp_allocator(), __x));
>> #endif
>> const size_type __elems_after = end() - __position;
>> pointer __old_finish(this->_M_impl._M_finish);
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Should reinterpret_cast be constexpr?
[not found] ` <CAFkJGRem_HFTtj93pyWVpvMtjSDLBU8u9nUVXshOTovsEjDaHw@mail.gmail.com>
@ 2020-12-29 1:19 ` Josh Marshall
0 siblings, 0 replies; 6+ messages in thread
From: Josh Marshall @ 2020-12-29 1:19 UTC (permalink / raw)
To: Jonathan Wakely, libstdc++
I think it is time for a check in. I have two tests failing, one related
to the code your snippet tries to fix and another which looks like a bug in
swapping elements.
https://gitlab.com/anadon/gcc
The lines of interest are
libstdc++-v3/testsuite/23_containers/vector/constexpr/swap_element.cc:67
(bug)
libstdc++-v3/testsuite/23_containers/vector/constexpr/swap_element.cc:130
(bug)
libstdc++-v3/testsuite/23_containers/vector/constexpr/function_insert.cc:35
(needs implementation)
Those are what I see as the current problems. Now might also be the time
to start examining the rest of the work. There's no way I kept to coding
standards when I'm this new.
On Mon, Dec 28, 2020 at 4:53 PM Josh Marshall <
joshua.r.marshall.1991@gmail.com> wrote:
> OK. Thanks for the freebie :)
>
> On Mon, Dec 28, 2020 at 4:25 PM Jonathan Wakely <jwakely.gcc@gmail.com>
> wrote:
>
>> On Mon, 28 Dec 2020 at 20:50, Josh Marshall
>> <joshua.r.marshall.1991@gmail.com> wrote:
>> >
>> > Is this in master? If so, I'll just merge it in.
>>
>> No, I just made a local change and haven't even committed it locally.
>>
>>
>> >
>> > On Mon, Dec 28, 2020 at 3:47 PM Jonathan Wakely <jwakely.gcc@gmail.com>
>> wrote:
>> >>
>> >> On Mon, 28 Dec 2020 at 20:33, Jonathan Wakely <jwakely.gcc@gmail.com>
>> wrote:
>> >> >
>> >> > On Mon, 28 Dec 2020 at 19:51, Josh Marshall via Libstdc++
>> >> > <libstdc++@gcc.gnu.org> wrote:
>> >> > >
>> >> > > Hello all,
>> >> > >
>> >> > > From the test error `libstdc++-v3/include/bits/stl_vector.h:1837:
>> error:
>> >> > > 'reinterpret_cast' is not a constant expression` and the
>> documentation at
>> >> > > https://en.cppreference.com/w/cpp/language/reinterpret_cast , is
>> >> > > `reinterpret_cast()` not constexpr in error?
>> >> >
>> >> > No, it's not an error. reinterpret_cast is not allowed in constant
>> expressions.
>> >>
>> >> This patch replaces the use of reinterpret_cast with an alternative
>> >> using C++20 features:
>> >>
>> >> diff --git a/libstdc++-v3/include/bits/vector.tcc
>> >> b/libstdc++-v3/include/bits/vector.tcc
>> >> index 27e63388feb..93b2ccc02e3 100644
>> >> --- a/libstdc++-v3/include/bits/vector.tcc
>> >> +++ b/libstdc++-v3/include/bits/vector.tcc
>> >> @@ -516,9 +516,13 @@ _GLIBCXX_BEGIN_NAMESPACE_CONTAINER
>> >> {
>> >> #if __cplusplus < 201103L
>> >> value_type __x_copy = __x;
>> >> -#else
>> >> +#elif ! __cpp_lib_make_obj_using_allocator
>> >> _Temporary_value __tmp(this, __x);
>> >> value_type& __x_copy = __tmp._M_val();
>> >> +#else
>> >> + value_type __x_copy = std::make_from_tuple<value_type>(
>> >> + std::uses_allocator_construction_args<value_type>(
>> >> + _M_get_Tp_allocator(), __x));
>> >> #endif
>> >> const size_type __elems_after = end() - __position;
>> >> pointer __old_finish(this->_M_impl._M_finish);
>>
>
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2020-12-29 1:19 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-12-28 19:51 Should reinterpret_cast be constexpr? Josh Marshall
2020-12-28 20:33 ` Jonathan Wakely
2020-12-28 20:47 ` Jonathan Wakely
2020-12-28 20:49 ` Josh Marshall
2020-12-28 21:24 ` Jonathan Wakely
[not found] ` <CAFkJGRem_HFTtj93pyWVpvMtjSDLBU8u9nUVXshOTovsEjDaHw@mail.gmail.com>
2020-12-29 1:19 ` Josh Marshall
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).