"false positive" in if_same_then_else when the branches differ by a comment

[nagisa]

I tried this code:

fn foo(a: bool, b: bool) -> Option<bool> {
    if a {
        Some(b)
    } else if b {
        // Some commentary about why this branch is important etc.
        None
    } else if !a {
        // Some more commentary about why this branch matters.
        None
    } else {
        Some(a)
    }
}

I expected to see this happen: no false positives!

Instead, this happened: if_same_then_else deny-by-default lint fired. While, yes, the branches are indeed "duplicate" at a logical level, they are not at a syntactic level. In this particular case the code ended up this way because it was important to document why we’re returning None if b and then None if !a, These comments could be rewritten to somehow combine them together, but there's probably nothing that’s clearer than just splitting code into a couple of the else if branches seen above.

Meta

• cargo clippy -V: clippy 0.0.212 (de521cb 2020-08-21)
• rustc -Vv: rustc 1.47.0-nightly (de521cb 2020-08-21)

---

(Knowing compilers, this is probably not super actionable, feel free to close if you aren’t too keen on adjusting behaviour here)

[nagisa]

(note that the conditionals aren’t actually as trivial as portrayed, but this was a reduction I could quickly extract from the actual example)

[tnielens]

This is not a false positive. The lint is designed so. If you consider having a separate branch more important here than applying the lint suggestion, you can annotate the fn definiction with #[allow(clippy::if_same_then_else)].

[nagisa]

Duplicate of #3770.