Detect and normalize line endings — \n (LF), \r\n (CRLF), and \r (CR). Zero dependencies and no_std. A Rust take on Node's detect-newline, plus the conversion and line-splitting you usually reach for next.

use detect_newline::{detect, dominant, to_lf, lines, Newline};

assert_eq!(detect("a\r\nb"), Some(Newline::CrLf));        // first ending found
assert_eq!(dominant("a\nb\nc\r\nd"), Some(Newline::Lf));  // most common
assert_eq!(to_lf("a\r\nb\rc"), "a\nb\nc");                // normalize
assert_eq!(lines("a\r\nb\rc").collect::<Vec<_>>(), ["a", "b", "c"]); // CR-aware

The Rust standard library's str::lines splits on \n/\r\n but not a lone \r, and there's no built-in way to detect a string's line-ending style or normalize a mixed string to one. This crate is the small, focused piece — pick a window of text apart by line ending, decide which one a file uses, and convert between them.

[dependencies]
detect-newline = "0.1"

• \r\n is always one line ending, never a \r followed by a \n — in detection, counting, normalization, and splitting alike.
• detect returns the first ending found; dominant returns the most frequent, breaking ties toward whichever style occurs first.
• normalize operates on character boundaries (safe for any UTF-8) and is idempotent.
• lines mirrors str::lines but also splits a lone \r; a single trailing line ending does not yield a final empty line.

#![no_std] (needs only alloc for String) with zero dependencies.

This project follows the all-contributors specification. Contributions of any kind are welcome — code, docs, bug reports, ideas, reviews! See the emoji key for how each contribution is recognized, and open a PR or issue to get involved.

Thanks goes to these wonderful people:

Tung Tran 💻 🚧

Licensed under either of Apache-2.0 or MIT at your option.