[Bug libstdc++/114925] New: include/bits/fs_path.h#L841 deprecation note suggests UB

[Bug libstdc++/114925] New: include/bits/fs_path.h#L841 deprecation note suggests UB

[fabian_kessler at gmx dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114925

            Bug ID: 114925
           Summary: include/bits/fs_path.h#L841 deprecation note suggests
                    UB
           Product: gcc
           Version: 15.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libstdc++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: fabian_kessler at gmx dot de
  Target Milestone: ---

```
template<typename _Source,
           typename _Require = __detail::_Path<_Source>,
           typename _CharT =
__detail::__value_type_is_char_or_char8_t<_Source>>
    _GLIBCXX20_DEPRECATED_SUGGEST("path((const char8_t*)&*source)")
    inline path
    u8path(const _Source& __source)
```

This is clearly undefined behavior, since it breaks aliasing rules. Even if
it's implementation vise ok, it might not be correct with other compilers.

[Bug libstdc++/114925] include/bits/fs_path.h#L841 deprecation note suggests UB

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114925

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I thought char8_t is still a character type so aliasing wise it falls under
that rule.

[Bug libstdc++/114925] include/bits/fs_path.h#L841 deprecation note suggests UB

[fabian_kessler at gmx dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114925

--- Comment #2 from fabian_kessler at gmx dot de <fabian_kessler at gmx dot de> ---
(In reply to Andrew Pinski from comment #1)
> I thought char8_t is still a character type so aliasing wise it falls under
> that rule.

Actually no. They are distinct types and only 4 types are allowed to alias
everything. That's char, signed/unsigned char and std::byte.

Consider the following function:

```
void cringe(char*& a, char8_t*& b){
    b = u8"Hello There!"
    print(a);
}
```

```
int main(){
   char8_t* obiwan = u8"Kenobi"; 
   cringe((char*&)obiwan, obiwan); // Might output "Kenobi" instead of "Hello
There!"   
}
```

That is, because the compiler is allowed to assume, that b can't alias a and no
change can't happen to a in that function.

[Bug libstdc++/114925] include/bits/fs_path.h#L841 deprecation note suggests UB

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114925

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2024-05-02
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW

--- Comment #3 from Jonathan Wakely <redi at gcc dot gnu.org> ---
The suggestion is wrong anyway because _Source doesn't have to be a contiguous
iterator, or any kind of iterator at all.