https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119754
Tomasz Kamiński <tkaminsk at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Ever confirmed|0 |1
CC| |tkaminsk at gcc dot gnu.org
Last reconfirmed| |2025-04-15
--- Comment #2 from Tomasz Kamiński <tkaminsk at gcc dot gnu.org> ---
The `clang` is right here, because allocator<int>::allocate() is specified to
start lifetime of array, but not it's elements
https://eel.is/c++draft/memory#allocator.members-5:
> This function starts the lifetime of the array object, but not that of any of
> the array elements.
However, our implementation just call `::operator new`, that starts lifetime of
both array and it's elements:
```
[[nodiscard,__gnu__::__always_inline__]]
constexpr _Tp*
allocate(size_t __n)
{
if (std::__is_constant_evaluated())
{
if (__builtin_mul_overflow(__n, sizeof(_Tp), &__n))
std::__throw_bad_array_new_length();
return static_cast<_Tp*>(::operator new(__n));
}
return __allocator_base<_Tp>::allocate(__n, 0);
}
```
We could call ~_Tp() for each trivially destructible types, but this will not
help with implicit lifetime aggregates with non-trivial destructor (we cannot
call it).
