fix: collapsible_match suggests ref/derefs when needed

[vishruth-thimmaiah]

Fixes: #14155

If an expression is borrowed/dereferenced, suggest using .as_ref() or .copied() respectively in the outer conditional statement.

changelog: [collapsible_match] suggests ref/dereferencing when needed.

[rustbot]

r? @llogiq

rustbot has assigned @llogiq.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[nk9]

Checking in here. It looks to me like the state is wrong now. Aren't we actually waiting on review?

[samueltardieu]

We are expecting both a review and a rebase. The bot settings have been changed a few days ago to let the PR in review state, so I'll reset it.

@rustbot review

[samueltardieu]

It looks like you ignore the fact that the reference might be mutable. The following code will suggest using .as_ref():

    let mut arr = ["a", "b", "c"];
    if let Some(mut last) = arr.last_mut() {
        match &mut last {
            &mut &mut "a" | &mut &mut "b" => {
                unimplemented!()
            },
            _ => (),
        }
    }

The suggestion might also be incorrect in case of raw references, where the types may not match (untested):

    const NULL_PTR: *const &'static str = std::ptr::null();
    if let Some(last) = arr.last() {
        match &raw const *last {
            NULL_PTR => unimplemented!(),
            _ => (),
        }
    }

[vishruth-thimmaiah]

I couldn't figure out any way to handle &raw refs, so I just disabled the lint for &raw refs. If you have any suggestions on how to handle them, I'd appreciate it!

Code has been updated.

[llogiq]

Sorry for the delay. Thanks for making clippy better! Can you open an issue for &raw handling? I don't think we should worry about the lint ignoring them, but someone may want to implement it anyway.

[samueltardieu]

@vishruth-thimmaiah Rebasing and reblessing the test output should be enough to make it mergeable again.

[rustbot]

This PR was rebased onto a different master commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.