New lint: `empty_expect` - detect `.expect("")` with empty message

What it does

Detects calls to .expect("") where the expect message is an empty string literal.

Why

When clippy::unwrap_used is denied, developers sometimes use .expect("") as a workaround — which provides no more context than .unwrap() on panic. The expect message should be meaningful.

Example

// Bad - should lint
let x: Option<i32> = Some(1);
let _ = x.expect("");

let v: Result<i32, &str> = Ok(1);
let _ = v.expect("");

// Good - no lint
let _ = x.expect("value should be present after initialization");
let _ = v.expect("parsing always succeeds for valid input");

Category

style or suspicious

Real-world motivation

In our codebase we deny unwrap_used to enforce better error context. Some developers bypass this by writing .expect(""), which defeats the purpose. Currently there is no Clippy lint to catch this — we resort to grep-based CI checks.

Start by locating existing Clippy lint implementations and their UI tests, then inspect how method calls and string literals are checked. Add coverage for .expect("") on both Option and Result, while confirming meaningful non-empty messages do not trigger; done means the new lint is categorized and documented consistently with neighboring lints.