Popular posts

What I don't like about Rust unicode is that string length and slicing happens at the byte level, not unicode glyph level. E.g. this panics: &"你好"[1..2]. It is understandable for Rust though, given its focus on performance.

Of course the correct way is also supported, and in typical Rust style it's explicit what the cost is, but few other langauges would consider this acceptable: "你好".chars().skip(1).take(1).collect::<String>()