String in a struct, lifetime

Question

I know what the difference between these 3 struct is - a lifetime of a

struct S1 {
    a: &'static str,
    b: int
}

struct S2<'aa> {
    a: &'aa str,
    b: int
}

struct S3 {
    a: String,
    b: int
}

fn main() {
    let s1 = S1 {a: "123", b: 123};
    let s2 = S2 {a: "123", b: 123};
    let s3 = S2 {a: "123".into_owned(), b: 123};
}

Could you show me a use case of the 1st, 2nd and 3rd, in other words, when is it better to use the 1st over 2nd and 3rd, when - 2nd over 1st and 3rd, etc? In the documentation there's no explanation.

Answer 1

S1: This only allows you to use string literals, or other strings with a static (i.e. they can never be deallocated) lifetime.

S2: This lets you use arbitrary string slices, provided they have an expressible lifetime. For example, you cannot return dynamic instances of S2 from an Iterator, because there's no way to express the lifetime involved. That said, this allows you to avoid unnecessary heap allocations.

S3: The most general, since it owns its contents, but requires heap allocation to use.