`#[expect(...)]` doen't catch lints from different `check_*` function

[Jarcho]

Some lints are emitted from a from a different from the area they are linted. e.g. clippy::ptr_arg is linting from check_body, but it can be allowed from the argument (fn foo(#[allow(clippy::ptr_arg)] x: &Vec<u32>) {}).

The following lints in clippy (and probably more) lint from different check_* functions:

• needless_borrow
• ref_binding_to_ref
• duplicate_mod
• multiple_inherent_impl
• macro_use_imports
• manual_non_exhaustive
• same_name_method
• async_yields_async
• await_holding_lock
• await_holding_refcell
• await_holding_invalid_type
• default_numeric_fallback
• ptr_arg
• redundant_clone
• logic_bug
• nonminimal_bool
• boxed_local
• implicit_return
• needless_return
• unnecessary_unwrap
• panicking_unwrap

Some of these don't actually work with allow at a narrower scope either, but some of them check to see if the lint has been allowed on the target item before emitting the lint.

cc #54503

[xFrednet]

@rustbot claim. I'll go through these lints one by one. The first two were caused by wrong emission functions on Clippy's side.

Todo List:

• needless_borrow
• ref_binding_to_ref
• same_name_method
• async_yields_async
• default_numeric_fallback
• ptr_arg
• logic_bug
• nonminimal_bool
• boxed_local
• implicit_return
• needless_return
• unnecessary_unwrap
• panicking_unwrap
• macro_use_imports
• manual_non_exhaustive (is_lint_allowed)

See: https://github.com/xFrednet/rust-clippy/tree/rust-97660-catch-emissions-with-expect for the current progress

• Requites new interface in rustc, to fulfill the expectation manually:

  • duplicate_mod (Early lint pass...)
  • multiple_inherent_impl (is_lint_allowed)

• Won't fix now:

  The expect attribute works on the body, but not on the guard expression. This is a result of the lint implementation and should therefore be fixed there. Other lint attributes are also ignored on the expr level

  • await_holding_lock
  • await_holding_refcell
  • await_holding_invalid_type

  The fix of these lints, will be tracked in Fix lint attributes/emission location for await_holding_* lints rust-clippy#9047

• Can't reproduce (But Test added):

  • redundant_clone

[xFrednet]

The next step for this issue is to adapt the linting interface of rustc to allow the manual fulfillment of expectations. Then I have to take a look at the last few lints, but they should all be covered by these changes.

@Jarcho, have you also looked at rustc lints or just at Clippy's lints? 🙃

[Jarcho]

It doesn't look like any of the rustc lints need to be changed. At least the behaviour of expect shouldn't be different than allow is.