new lint: borrow_deref_ref

[lengyijun]

changelog: [`borrow_deref_ref`]

Related pr: #6837 #7577
@Jarcho Could you please give a review?

cargo lintcheck gives no false negative (but tested crates are out-of-date).

TODO:

1. Not sure the name. deref_on_immutable_ref or some others?

[rust-highfive]

r? @Manishearth

(rust-highfive has picked a reviewer for you, use r? to override)

[Manishearth]

r? @Jarcho

[flip1995]

Some general comments. I wasn't involved in the previous discussions about this lint, so I don't really know how we wanted to move forward with this lint.

[camsteffen]

This lint is very similar to deref_addrof but with the operators swapped. How about the name addrof_deref_ref - &*<ref-typed-expr>.

[camsteffen]

Though we probably should say "borrow" instead of "addrof" since that's what the docs say - so maybe borrow_deref_ref.

[lengyijun]

Conclusion and current status:

We lint on:

1. let x = &* (&T)
2. let x = &mut &* foo()
3. *&* (&T) if DEREF_ADDROF is silent

We don't lint on:

1. &&* (&T)
2. &mut &* (&T)

[flip1995]

Here's another interesting case, I stumbled upon working on rustc today: Playground

struct S<'a> {
    a: &'a mut u32,
}

fn f(s: &S<'_>) {
    let x = &*s.a;
    // let x = s.a; // <- errors, since &mut must be moved and cannot be copied
    println!("{:?}", x);
}

[bors]

☔ The latest upstream changes (presumably #8025) made this pull request unmergeable. Please resolve the merge conflicts.

[bors]

☔ The latest upstream changes (presumably #8042) made this pull request unmergeable. Please resolve the merge conflicts.

[bors]

☔ The latest upstream changes (presumably #8127) made this pull request unmergeable. Please resolve the merge conflicts.

[Jarcho]

Sorry for the late review. December's been very busy.

Both help messages should be suggestions instead.

[lengyijun]

Both help messages should be suggestions instead.

Suggestion looks much better!

[bors]

☔ The latest upstream changes (presumably #8621) made this pull request unmergeable. Please resolve the merge conflicts.

[xFrednet]

Hey @Jarcho, could you leave a normal comment (not a review) on this PR, so I can assign it to you?

r? @Manishearth could you afterwards do the final review for this PR? 🙃

[Jarcho]

Can do.

[Manishearth]

@bors delegate=Jarcho

but yeah sure!

[bors]

✌️ @Jarcho can now approve this pull request

[Jarcho]

Everything looks good. Just have the changes to rustc_tools_util/LICENSE-APACHE and rustc_tools_util/LICENSE-MIT to revert. Feel free to squash while you're at it.

[lengyijun]

Run cargo test --features deny-warnings,internal locally emit no error,
but CI give some errors.

[xFrednet]

On master, we've added a new test, that ensures that lints with automatically applicable suggestions also have a .fixed file. It looks like the CI failed due to this. It might be good to rebase on master again :)

Even though, I'm surprised that this fails in the CI for you

[Jarcho]

You're going to need to split the test file into two: borrow_deref_ref.rs and borrow_deref_ref_unfixable.rs. Move everything that produces two suggestions over to the unfixable one and then add an exemption to RUSTFIX_COVERAGE_KNOWN_EXCEPTIONS in compile-test.rs for the unfixable test.

[Jarcho]

Thanks for keeping up with everything!

@bors r+

[bors]

📌 Commit 202fdb9 has been approved by Jarcho

[bors]

⌛ Testing commit 202fdb9 with merge c4c413b...

[bors]

☀️ Test successful - checks-action_dev_test, checks-action_remark_test, checks-action_test
Approved by: Jarcho
Pushing c4c413b to master...