new lint: method_without_self_relation

[andrewgazelka]

Adds a new restriction lint that detects methods in impl blocks that have no relationship to Self, neither through parameters, return types, nor receiver. These are better expressed as standalone functions.

Motivation

Methods that don't reference Self anywhere in their signature are not truly "methods" but rather associated functions that happen to be namespaced under a type. Moving them to standalone functions can improve code modularity and discoverability.

Implementation Details

• Category: restriction - intentionally conservative since there are valid reasons to keep functions in impl blocks (helpers, private implementation details, namespace organization)
• Detection: Recursively checks function signatures for Self references in: receiver and return types

Examples

Triggers lint:

impl Calculator {
    fn add(a: i32, b: i32) -> i32 {  // No Self relation
        a + b
    }
}

Does not trigger:

impl Calculator {
    // Direct Self references
    fn new(precision: u32) -> Self { ... }  // Returns Self
    fn from_self(other: Self) -> u32 { ... }  // Takes Self as parameter

    // Self in generics (symmetric - parameters and return types)
    fn from_option(opt: Option<Self>) -> u32 { ... }  // Self in parameter generic
    fn maybe_new(precision: u32) -> Option<Self> { ... }  // Self in return type generic
    fn try_new(precision: u32) -> Result<Self, String> { ... }  // Self in Result
    fn tuple_with_self(x: i32) -> (i32, Self) { ... }  // Self in tuple return

    // Receivers
    fn get_precision(&self) -> u32 { ... }  // Has &self receiver
    fn consume(self) -> u32 { ... }  // Has self receiver
}

changelog: [method_without_self_relation]: new restriction lint to detect methods with no relationship to Self

[github-actions]

Lintcheck changes for 17410f9

Lint	Added	Removed	Changed
clippy::method_without_self_relation	354	0	0

This comment will be updated if you push new changes

[rustbot]

r? @y21

rustbot has assigned @y21.
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

[y21]

It looks like this triggers for methods in trait impls - that seems like a FP? E.g. there's a lintcheck warning on https://docs.rs/ahash/0.8.11/src/ahash/specialize.rs.html#95-100

[andrewgazelka]

It looks like this triggers for methods in trait impls - that seems like a FP? E.g. there's a lintcheck warning on docs.rs/ahash/0.8.11/src/ahash/specialize.rs.html#95-100

38587b0

[andrewgazelka]

It looks like this triggers for methods in trait impls - that seems like a FP? E.g. there's a lintcheck warning on docs.rs/ahash/0.8.11/src/ahash/specialize.rs.html#95-100

any updates?

[andrewgazelka]

Hey @y21 👋

Just wanted to clarify - the trait impl false positive you flagged was fixed in commit 38587b0 ("do not trigger for trait impls"). The lint now explicitly skips trait implementations:

// Don't lint trait implementations - these methods are defined by the trait
if implements_trait {
    return;
}

The lintcheck was re-run after this fix and shows 354 warnings on real-world crates, none of which are from trait impls anymore.

Just rebased onto latest master. Let me know if there's anything else needed for review!

[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.

[andrewgazelka]

@rustbot ready

[rustbot]

☔ The latest upstream changes (possibly #15979) made this pull request unmergeable. Please resolve the merge conflicts.