Make Box drop through Drop trait

[beepster4096]

This PR makes Box drop through its implementation of the Drop trait, instead of by being special-cased in the compiler.

Blocked on #93028

[rust-highfive]

r? @m-ou-se

(rust-highfive has picked a reviewer for you, use r? to override)

[rust-log-analyzer]

The job mingw-check failed! Check out the build log: (web)(plain)

Click to see the possible cause of the failure (guessed by this bot)

 Checking alloc v0.0.0 (/checkout/library/alloc)
Checking cfg-if v0.1.10
Checking adler v0.2.3
Checking rustc-demangle v0.1.21
error[E0367]: `Drop` impl requires `A: core::alloc::Allocator` but the struct it is implemented for does not
|
|
1176 | A: ~const Allocator,
|
note: the implementor must specify the same requirement
--> library/alloc/src/boxed.rs:175:1
|
|
175 | / pub struct Box<
176 | | T: ?Sized,
177 | | #[unstable(feature = "allocator_api", issue = "32838")] A: Allocator = Global,
178 | | >(Unique<T>, A);
For more information about this error, try `rustc --explain E0367`.
error: could not compile `alloc` due to previous error
Build completed unsuccessfully in 0:01:21

[beepster4096]

Alright good, now that CI is failing in the way it should, all we have to do is wait for #93028 to merge.

[nagisa]

I suspect this will run into problems with ensuring that the memory is indeed freed in case of incomplete initialization as part of the box syntax or in case of panics during drop.

Would be a good idea to verify there are tests for these behaviours and if not – to implement some.

[beepster4096]

Actually, this is still going to need some compiler magic for dropping T because of some of Box's other magic.

[beepster4096]

Honestly, don't think this is that useful because of that. Not gonna bother finishing this.