[Bug libstdc++/112498] New: std::is_convertible<float, _Float16>::value returns false

[Bug libstdc++/112498] New: std::is_convertible<float, _Float16>::value returns false

[josopait at goopax dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112498

            Bug ID: 112498
           Summary: std::is_convertible<float, _Float16>::value returns
                    false
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libstdc++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: josopait at goopax dot com
  Target Milestone: ---

If I am not mistaken, this should succeed:


#include <type_traits>

int main()
{
  static_assert(std::is_convertible<double, float>::value);   // ok
  static_assert(std::is_convertible<float, _Float16>::value); // fails
}

[Bug c++/112498] std::is_convertible<float, _Float16>::value returns false

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112498

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |INVALID
             Status|UNCONFIRMED                 |RESOLVED
          Component|libstdc++                   |c++
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=111842

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
No.
```
_Float16 f(float a)
{
  return a;
}
```

is invalid C++23 code.
<source>:3:10: warning: converting to '_Float16' from 'float' with greater
conversion rank

Maybe we should add this to the FAQ because we get a question about implicit
conversion from _Float16 to float a few times now.


See
https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2020/p1467r4.html#implicit

Which is the C++ paper which adds the extended floating point types and GCC
implements it full. Just as an extension allows the conversion (with a warning)
but since we cannot have different behavior for sfinae and concepts with
-pedantic/-pedantic-errors std::is_convertible needs to return false.

[Bug c++/112498] std::is_convertible<float, _Float16>::value returns false

[josopait at goopax dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112498

--- Comment #2 from Ingo Josopait <josopait at goopax dot com> ---
Yes, you are right. It is confusing, though. Maybe the warning message should
mention that it is actually invalid code. It looked more like a suggestion to
me.

[Bug c++/112498] std::is_convertible<float, _Float16>::value returns false

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112498

--- Comment #3 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Maybe we could add "is not allowed in ISO C++" or similar to the warning.

[Bug c++/112498] std::is_convertible<float, _Float16>::value returns false

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112498

--- Comment #4 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jonathan Wakely <redi@gcc.gnu.org>:

https://gcc.gnu.org/g:69d69865a792a93cce2905617c53913769d0f260

commit r14-5436-g69d69865a792a93cce2905617c53913769d0f260
Author: Jonathan Wakely <jwakely@redhat.com>
Date:   Mon Nov 13 12:03:31 2023 +0000

    c++: Link extended FP conversion pedwarns to -Wnarrowing [PR111842]

    Several users have been confused by the status of these warnings,
    which can be misunderstood as "this might not be what you want",
    rather than diagnostics required by the C++ standard. Add the text "ISO
    C++ does not allow" to make this clear.

    Also link them to -Wnarrowing so that they can be disabled or promoted
    to errors independently of other pedwarns.

            PR c++/111842
            PR c++/112498

    gcc/cp/ChangeLog:

            * call.cc (convert_like_internal): Use OPT_Wnarrowing for
            pedwarns about illformed conversions involving extended
            floating-point types. Clarify that ISO C++ requires these
            diagnostics.

    gcc/testsuite/ChangeLog:

            * g++.dg/cpp23/ext-floating16.C: New test.
            * g++.dg/cpp23/ext-floating17.C: New test.

[Bug c++/112498] std::is_convertible<float, _Float16>::value returns false

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112498

--- Comment #5 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Jonathan Wakely from comment #3)
> Maybe we could add "is not allowed in ISO C++" or similar to the warning.

This has been done for GCC 14:

fp.cc:2:20: warning: ISO C++ does not allow converting to 'std::float16_t' {aka
'_Float16'} from 'float' with greater conversion rank [-Wnarrowing]