Replace #[allow] with #[expect] in Clippy

[xFrednet]

Hey @rust-lang/clippy, @Alexendoo, @dswij, I'm currently working on the expect attribute as defined in Rust RFC 2383. With that, an #[allow] attribute can be replaced with a #[expect] attribute that suppresses the lint, but also emits a warning, if the lint isn't emitted in the expected scope.

With this PR I would like to test the attribute on a project scale and Clippy obviously came to mind. This PR replaces (almost) all #[allow] attributes in clippy_utils and clippy_lints with the #[expect] attribute. I was also able to remove some allows since, the related FPs have been fixed 🎉.

My question is now, are there any concerns regarding this? It's still okay to add normal #[allow] attributes, I see the need to nit-pick about that in new PRs, unless it's actually a FP. Also, I would not recommend using #[expect] in tests, as changes to a lint could the trigger the expect attribute in other files.

Additionally, I've noticed that Clippy has a bunch of #[allow(clippy::too_many_lines)] attributes. Should we maybe allow the lint all together or increase the threshold setting? To me, it seems like we mostly just ignore it in our code. 😅 🙃

---

changelog: none

r? @flip1995 (I've requested you for now, since you're also helping with reviewing the expect implementation. You are welcome to delegate this PR, even if it should be a simple review 🙃 )

[xFrednet]

Also, just to make sure

@bors try

[bors]

⌛ Trying commit 66b5c71 with merge 88af5f2...

[flip1995]

I think this is a good move. This will also give us tests for FPs and we won't miss fixes to them anymore as in the write.rs file.

Let's wait if other people have concerns. But I don't really see what would speak against this.

[flip1995]

Regarding the too_many_lines lint: I would keep it enabled by default with the same threshold. I think most places we allow this lint are in functions that just contain one huge match statement.

If this is added in a new PR, we might want to see what we can do to simplify/split up the function.

[bors]

☀️ Try build successful - checks-action_dev_test, checks-action_remark_test, checks-action_test
Build commit: 88af5f2 (88af5f29dbc892f9fe0637ff7ea3a1c74bef5f70)

[xFrednet]

I forgot to add the expect changes in clippy_utils, now they should be added 🙃

[Manishearth]

I think we should do this 👍

[Jarcho]

Unrelated to the PR itself, but relevant for clippy. What happens to is_lint_allowed when using #[expect(...)]

[xFrednet]

Interesting question, the expect implementation requires the lint to be emitted like it would be for a warning. Therefore, is_lint_allowed will return false. This behavior is also documented in the new lint level documentation 🙃

[xFrednet]

It looks there are no concerns so far. @flip1995, I've pushed an update with the expects for clippy_utils since your review. Would you mind reviewing it and r+ if everything looks good? 🙃

[flip1995]

@bors r+

Thanks!

[bors]

📌 Commit 03960eb has been approved by flip1995

[bors]

⌛ Testing commit 03960eb with merge aa03344...

[bors]

☀️ Test successful - checks-action_dev_test, checks-action_remark_test, checks-action_test
Approved by: flip1995
Pushing aa03344 to master...