Check if dropping an expression may have indirect side-effects

[samueltardieu]

It is not enough to check if an expression type implements Drop to determine whether it can have a significant side-effect.

Also, add tests for unchecked cases which were explicitly handled in the code, such as checking for side effect in a struct's fields, or in its base expression.

Fix #14592

changelog: [no_effect_underscore_binding]: do not propose to remove the assignment of an object if dropping it might indirectly have a visible side effect

[rustbot]

r? @blyxyas

rustbot has assigned @blyxyas.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[blyxyas]

I'm not being able to test this, but are we also interested in replacing unnecessary_operation's has_drop uses into this new function?

[samueltardieu]

I'm not being able to test this, but are we also interested in replacing unnecessary_operation's has_drop uses into this new function?

No, it would lose the ability to suggest replacing main()'s content in this code

struct WithDrop;
impl Drop for WithDrop {
    fn drop(&mut self) {
        todo!()
    }
}

struct WithoutDrop { inner: WithDrop }

fn main() {
    (WithoutDrop { inner: WithDrop });
}

by

    WithDrop;

as it is currently doing.

[blyxyas]

LGTM, thanks! ❤️ =^w^=

[samueltardieu]

@blyxyas Did you forget to press merge by chance?

[blyxyas]

Yeah, I approved it before getting on the bus and suffered from a brain wipe (on the good side, I now have a draft for the blog post)