Fix [use_self] false negative with on struct and tuple struct patterns

[botahamec]

fixes #8845

changelog: Triggered the warning for [`use_self`] on TupleStruct and Struct patterns, whereas currently it's only triggered for Path patterns

[rust-highfive]

r? @xFrednet

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

[xFrednet]

r? @Alexendoo (It would be cool, if you could leave a comment so that I can assign this PR to you :))

[Alexendoo]

👋

[Alexendoo]

Using the variants directly trigger the lint, suggesting to replace them all with Self:

fn imported_variants(&self) -> u8 {
    use Something::*;

    match self {
        Num(n) => *n,
        TupleNums(n, _m) => *n,
        StructNums { one, two: _ } => *one,
    }
}

[Alexendoo]

Seems that's uncovering an existing bug, the first path segment is picked unconditionally

enum Enum {
    A,
}

impl Enum {
    fn f(self) -> usize {
        match self {
            crate::Enum::A => 1,
        }
    }
}

warning: unnecessary structure name repetition
  --> t.rs:13:13
   |
13 |             crate::Enum::A => 1,
   |             ^^^^^ help: use the applicable keyword: `Self`
   |

You could check the res field of the PathSegments to pick the right one

[botahamec]

This was tricky. Using the res field on the PathSegments doesn't seem to work. In the playground, I found that most of the res fields had errors, so I ended up using a workaround similar to what's done in the check_expr method. Luckily, the lint_path_to_variant function already exists. But, it's a bit hacky, and I did find that the name of the enum is correctly shown in the res fields. I made a version of the function which uses the res field, and is a bit less hacky, but it's also a bit less performant. I doubt anyone would notice the difference, but I left it unchanged for now.

[Alexendoo]

Thanks! Could you also squash the commits and then I think it'll be good to go

[botahamec]

@Alexendoo Should be good now

[xFrednet]

@Alexendoo when you're happy with this PR you can r+ it directly. And we no longer need a delegate for that 🥳

[Alexendoo]

@bors r+

[bors]

@Alexendoo: 🔑 Insufficient privileges: Not in reviewers

[Alexendoo]

Well then. Guess the teams PR needs to happen for that yet 😄

[xFrednet]

Talking in bors voice

Thank you for the approval

@bors r=Alexendoo

[bors]

📌 Commit 2aa4569 has been approved by Alexendoo

[bors]

⌛ Testing commit 2aa4569 with merge d9f4978...

[bors]

☀️ Test successful - checks-action_dev_test, checks-action_remark_test, checks-action_test
Approved by: Alexendoo
Pushing d9f4978 to master...