Lint for generalizeable captures

[est31]

What it does

Take this code:

fn foo(v: Option<char>) -> u32 {
    match v {
        Some(w @ 'a' | w @ 'c' | w @ 'd') => w as u32,
        Some(_) => 1,
        None => 0,
    }
}

The binding w of the first arm's pattern is repeated three times. One can simplify the pattern as follows:

fn foo(v: Option<char>) -> u32 {
    match v {
        Some(w @ ('a' | 'c' | 'd')) => w as u32,
        Some(_) => 1,
        None => 0,
    }
}

Lint Name

repetitive_capture

Category

style

Advantage

This saves repetition of the binding (less chars) and makes the code more maintainable as renames only affect one site instead of multiple

Drawbacks

It adds one layer of ()s, but this should be bearable

Example

provided above

[est31]

See also rust-lang/rust#109489 where I found instances of this in rustc.

[blyxyas]

@rustbot label: +good-first-issue

[Almugra]

@rustbot claim

[Almugra]

I think unnested_or_patterns should already handle this, or is it different because unnested_or_patterns is pedantic? If that is the case, I would like to claim it again.

[est31]

Hmm good point about unnested_or_patterns. I have tried enabling all the clippy lints before I filed the issue but it didn't show up, not sure what went wrong back then.

error: unnested or-patterns
 --> src/main.rs:5:9
  |
5 |         Some(w @ 'a' | w @ 'c' | w @ 'd') => w as u32,
  |         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  |
  = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#unnested_or_patterns
note: the lint level is defined here
 --> src/main.rs:1:8
  |
1 | #[deny(clippy::pedantic)]
  |        ^^^^^^^^^^^^^^^^
  = note: `#[deny(clippy::unnested_or_patterns)]` implied by `#[deny(clippy::pedantic)]`
help: nest the patterns
  |
5 |         Some(w @ ('a' | 'c' | 'd')) => w as u32,
  |         ~~~~~~~~~~~~~~~~~~~~~~~~~~~

Closing as it's already implemented in clippy.