[Bug c++/99546] New: Weird return value of C++20 requires expression

[Bug c++/99546] New: Weird return value of C++20 requires expression

[hewillk at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99546

            Bug ID: 99546
           Summary: Weird return value of C++20 requires expression
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: hewillk at gmail dot com
  Target Milestone: ---

gcc-trunk accepts following code:

int main() {
  constexpr auto b = requires { []{}; };
  static_assert(b);
  static_assert(!b);
}

(godbolt: https://godbolt.org/z/4K5vq8)

[Bug c++/99546] Weird return value of C++20 requires expression

[hewillk at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99546

--- Comment #1 from 康桓瑋 <hewillk at gmail dot com> ---
You can see more weird cases in this: 
https://stackoverflow.com/questions/66578966/weird-return-value-of-c20-requires-expression.

[Bug c++/99546] Weird return value of C++20 requires expression

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99546

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
           Keywords|wrong-code                  |accepts-invalid
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2021-03-11

--- Comment #2 from Jonathan Wakely <redi at gcc dot gnu.org> ---
I think the requires expression should be rejected as ill-formed.

[Bug c++/99546] Weird return value of C++20 requires expression

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99546

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|accepts-invalid             |wrong-code

--- Comment #3 from Jonathan Wakely <redi at gcc dot gnu.org> ---
My mistake, the requires-expression is valid, so it is just a wrong-code big as
you originally said.

[Bug c++/99546] Weird return value of C++20 requires expression

[hewillk at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99546

--- Comment #4 from 康桓瑋 <hewillk at gmail dot com> ---
same form, but this will trigger ICE:

====

struct S{
  constexpr static auto s = requires { []; };
};

====

<source>:1:8: internal compiler error: tree check: expected class 'type', have
'exceptional' (error_mark) in override_type_for_decl_p, at dwarf2out.c:23786
    1 | struct S{
      |        ^
0x1cfa229 internal_error(char const*, ...)
        ???:0
0x67b5e5 tree_class_check_failed(tree_node const*, tree_code_class, char
const*, int, char const*)
        ???:0
0xf3666f rest_of_type_compilation(tree_node*, int)
        ???:0
0x70c969 finish_struct_1(tree_node*)
        ???:0
0x70e754 finish_struct(tree_node*, tree_node*)
        ???:0
0x8e121d c_parse_file()
        ???:0
0xa5f812 c_common_parse_file()
        ???:0
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.

[Bug c++/99546] Weird return value of C++20 requires expression

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99546

--- Comment #5 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to 康桓瑋 from comment #4)
> same form, but this will trigger ICE:

That's a completely different bug. I've opened PR 106976

[Bug c++/99546] Weird return value of C++20 requires expression

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99546

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|2021-08-16 00:00:00         |2022-9-20

--- Comment #6 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to 康桓瑋 from comment #1)
> You can see more weird cases in this: 
> https://stackoverflow.com/questions/66578966/weird-return-value-of-c20-
> requires-expression.

Copying that here instead of linking to impermanent SO comments:

template <bool b>
void foo() {
    static_assert(b);
}

int main() {
  constexpr bool b = requires { []{}; };
  foo<!b>();
  foo<b>();
}

The bool appears to not be a constant expression:

x.C: In function 'int main()':
x.C:8:10: error: no matching function for call to 'foo<(!(bool)b)>()'
    8 |   foo<!b>();
      |   ~~~~~~~^~
x.C:2:6: note: candidate: 'template<bool b> void foo()'
    2 | void foo() {
      |      ^~~
x.C:2:6: note:   template argument deduction/substitution failed:
x.C:9:9: error: no matching function for call to 'foo<b>()'
    9 |   foo<b>();
      |   ~~~~~~^~
x.C:2:6: note: candidate: 'template<bool b> void foo()'
    2 | void foo() {
      |      ^~~
x.C:2:6: note:   template argument deduction/substitution failed:

[Bug c++/99546] Weird return value of C++20 requires expression

[gieseanw+gcc at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99546

Andrew Giese <gieseanw+gcc at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |gieseanw+gcc at gmail dot com

--- Comment #7 from Andrew Giese <gieseanw+gcc at gmail dot com> ---
This bug in the OP appears to have been fixed as early as gcc 11.1.

For reference, I'm referring to the following code:

int main() {
  constexpr auto b = requires { []{}; };
  static_assert(b);
  static_assert(!b);
}

https://godbolt.org/z/WzzGMxcz6


Note that the code listed in Jonathan's comment above still fails in 14.0.1
(https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99546#c6)

[Bug c++/99546] Weird return value of C++20 requires expression

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99546

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|2022-09-20 00:00:00         |2024-2-15

--- Comment #8 from Jonathan Wakely <redi at gcc dot gnu.org> ---
The OP's example is supposed to be ill-formed though, it shouldn't compile
successfully. So I don't think it's fixed.

[Bug c++/99546] Weird return value of C++20 requires expression

[gieseanw+gcc at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99546

--- Comment #9 from Andrew Giese <gieseanw+gcc at gmail dot com> ---
(In reply to Jonathan Wakely from comment #8)
> The OP's example is supposed to be ill-formed though, it shouldn't compile
> successfully. So I don't think it's fixed.

That's my mistake, sorry. Yes the second static assertion should definitely
fail.