Docs: clarify that certain intrinsics are not unsafe

[bstrie]

As determined by the hardcoded list at

rust/compiler/rustc_typeck/src/check/intrinsic.rs

Lines 59 to 92 in 003b8ea

	sym::abort
	| sym::size_of
	| sym::min_align_of
	| sym::needs_drop
	| sym::caller_location
	| sym::add_with_overflow
	| sym::sub_with_overflow
	| sym::mul_with_overflow
	| sym::wrapping_add
	| sym::wrapping_sub
	| sym::wrapping_mul
	| sym::saturating_add
	| sym::saturating_sub
	| sym::rotate_left
	| sym::rotate_right
	| sym::ctpop
	| sym::ctlz
	| sym::cttz
	| sym::bswap
	| sym::bitreverse
	| sym::discriminant_value
	| sym::type_id
	| sym::likely
	| sym::unlikely
	| sym::ptr_guaranteed_eq
	| sym::ptr_guaranteed_ne
	| sym::minnumf32
	| sym::minnumf64
	| sym::maxnumf32
	| sym::rustc_peek
	| sym::maxnumf64
	| sym::type_name
	| sym::forget
	| sym::variant_count => hir::Unsafety::Normal,

[rust-highfive]

r? @kennytm

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

[RalfJung]

Maybe add a comment in rustc_typeck/src/check/intrinsic.rs telling people to also update the docs when adding more items to that list?

[lqd]

Btw: #86327 will fix the rustdoc issue which made these intrinsics' signatures appear unsafe to call in the libcore/libstd docs.

[bstrie]

Btw: #86327 will fix the rustdoc issue which made these intrinsics' signatures appear unsafe to call in the libcore/libstd docs.

Great to hear that the rustdoc bug is fixed. I still think that this PR would be useful, since the idea of any safe function declared inside an extern block is surprising, and users used to the normal semantics of extern blocks could just as easily assume that the lack of a visible unsafe in the docs is a bug in rustdoc. :P

[JohnTitor]

Looks reasonable to me! It's a bit unfortunate that we have to write a lot of the same comments but it should be fine for now.

[JohnTitor]

r? @JohnTitor@bors r+ rollup

[bors]

📌 Commit 16168dd has been approved by JohnTitor