vec_init_then_push incorrectly fires when the `push()` is in a `loop` body

[duci9y]

Lint name: vec_init_then_push

I tried this code:

let mut v = Vec::new();
loop {
    match *s {
        ...
        SExp::Cons(ref a, ref b) => {
            v.push(a);
        }
        ...
    }
}

I expected to see this happen: The lint to not fire, since this is a valid use case for init then push.

Instead, this happened: The lint fired.

378 | /         let mut v = Vec::new();
379 | |         loop {
380 | |             match *s {
381 | |                 ...
382 | |                 SExp::Cons(ref a, ref b) => {
383 | |                     v.push(a);
    | |______________________________^ help: consider using the `vec![]` macro: `let mut v = vec![..];`

Meta

• cargo clippy -V: clippy 0.1.51 (2fd73fab 2021-03-23)
• rustc -Vv:

rustc 1.51.0 (2fd73fabe 2021-03-23)
binary: rustc
commit-hash: 2fd73fabe469357a12c2c974c140f67e7cdd76d0
commit-date: 2021-03-23
host: x86_64-apple-darwin
release: 1.51.0
LLVM version: 11.0.1

[camsteffen]

This should be fixed by #6697. Can you provide a minimal reproducible example?

[duci9y]

I spent about 15 minutes trying to get a repro going but couldn't trigger the false positive. I'll try again later or see what I can gather about the context.

[TyPR124]

I came across this, and managed to simplify syntax down to this. I can reproduce locally clippy 0.1.51 (2fd73fa 2021-03-23), but not on play.rust-lang.org 2021-04-27 727d101 (as of writing, true for both this example and my real code).

#[allow(clippy::while_let_loop)]
#[allow(clippy::never_loop)]
#[allow(clippy::match_single_binding)]
fn foo() {
    let mut v = vec![];
    loop {
        match () {
            () => {
                v.push(());
                return;
            }
        }
    }
}

[camsteffen]

I can reproduce locally clippy 0.1.51 (2fd73fa 2021-03-23)

That version predates #6697. Your example no longer lints.

[TyPR124]

That's why I mentioned play.rust-lang.org as well, as I assume that version included that fix. Is there something that prevents that example from being a MCVE?

[camsteffen]

We need an example that is a false positive in the latest version, or else there is no issue here.

[TyPR124]

But the issue was opened for clippy 0.1.51 (2fd73fab 2021-03-23), the same version I was using.

or else there is no issue here.

Yes, my example leads me to believe this issue has been resolved.