is digits instead of digits range contains

[leonardo-m]

What it does

It suggests to replace a coding pattern of using range + "contains" with a more specific function.

Lint Name

is_ascii_digit_instead_contains

Category

style

Advantage

Less bug-prone, shorter

Drawbacks

No response

Example

fn foo1(d: u8) -> bool {
    (b'0' ..= b'9').contains(&d)
}

Could be written as:

fn foo2(d: u8) -> bool {
    d.is_ascii_digit()
}

[koka831]

It is supported in https://rust-lang.github.io/rust-clippy/master/#manual_is_ascii_check :)

[leonardo-m]

Are you sure that lint fires for that foo1 function? I've tried with this and still I don't see anything.

#![warn(clippy::all)]
#![warn(clippy::nursery)]
#![warn(clippy::pedantic)]

[koka831]

Ah, I'm sorry I misunderstood it. The lint only supports for matches! macro and does not for contains currently.

[xFrednet]

This can be an addition to the lint logic. :)

@rustbot label +good-first-issue

[naosense]

@rustbot claim

Any hints where I can start?😀

[koka831]

Hi @naosense, you can add check for contains method in is_manual_ascii_check lint.
Currently it looks for matches! macro only.

rust-clippy/clippy_lints/src/manual_is_ascii_check.rs

Lines 78 to 80 in 1a96571

	let Some(macro_call) = root_macro_call(expr.span) else { return };
	if is_matches_macro(cx, macro_call.def_id) {

You may find further informations in the book https://doc.rust-lang.org/nightly/clippy/development/index.html :)

[xFrednet]

Hey @naosense you can start, by checking for <expr-1>.contains(<expr-2>) calls. If <expr-2> is a character you can then check if <expr-1> is a specific range.

When starting on a new lint I often first look at the HIR with the #[clippy::dump] attribute. Try running Clippy on this playground. Then you "just" need to write the logic to match the printed HIR tree.

As always you're welcome to reach out if you want more guidance. @koka831 already correctly suggested, adding this in the rust-clippy/clippy_lints/src/manual_is_ascii_check.rs file. Mostly to keep all lint logic in one place

[naosense]

Does anyone use intellj, I configured it as instructed, but it didn't look good, see #9969. Anything wrong?

[Niki4tap]

@xFrednet this now can be closed since #10053 landed?

[llogiq]

Agreed. Let's close this.