Suggestion of `collect_push_then_iter` for iterators

[profetia]

What it does

When an iterator is collected into vector, which is then push()-ed into right after and then turned backed into iterator to do other things, it should be rewrite into a chain() instead.

Advantage

• Make the code more readable
• Remove redudant allocations for better performance

Drawbacks

• Not sure if the compiler is smart enough to optimize it out

Example

fn collect_push_then_iter(iter: impl Iterator<Item = i32>) -> Vec<i32> {
    let mut v = iter.collect::<Vec<_>>();
    v.push(1);
    v = v.into_iter().map(|x| x + 1).collect();
    v
}

Could be written as:

fn collect_push_then_iter(iter: impl Iterator<Item = i32>) -> Vec<i32> {
    iter.chain(std::iter::once(1)).map(|x| x + 1).collect()
}

[profetia]

I saw this when working on Clippy's source code, not sure if this is a common mistake

[samueltardieu]

Note that using .chain(std::iter::once(1)) is overkill in this case, as .chain([1]) will do the job (or .chain(Some(1))).