Must use without note

[skatromb]

PR addresses proposal for adding a lint that warns about #[must_use] attribute that lacks a note.

---

Request for a help

I feel I need some opinions here. I stuck on dogfood tests after running cargo test --test dogfood.

The issue is that I received a lot of lint warnings regarding the #[must_use] attribute without a note (which was what we wanted). However, the more I looked into them, the more I realized that most of the attributes do not make much sense. For example, they are set on getters and formatters. As I started deleting this attribute from these functions, I immediately received a "must_use_candidate" lint warning. One could argue that this is correct because if you call a "pure" function and don't use its result, you've done something wrong, so they're perfect candidates for must_use. On the other hand, setting this attribute on almost every function bloats the code and devalues lint. Adding a note "just because the function is pure" on must_use worsens the situation even more.

I ended up in a situation where I couldn't remove the "must_use" attribute without turning off the "must_use_candidate" lint. On the other hand, I feel it should be up to the developer to decide whether to mark a function as #[must_use] or not, and it shouldn't be so formal. A computational-heavy function or the necessity to handle important state could be good reasons, but they shouldn't be the only formal reasons for function purity. As far as I can tell, the only situation prevented by "must_use_candidate" is calling a function without the "let = must_use_func()" assignment: must_use_func(). Other cases, such as assigning to a variable and not using it, are covered by other lints.

I also found a standard library developers guide: when to add #[must_use], which suggests "The #[must_use] attribute can be applied to types or functions when failing to explicitly consider them or their output is almost certainly a bug."

So what do you think, should I fix all must_use attributes adding some formal note, or should we reconsider must_use_candidate lint?

---

changelog: [must_use_without_note]: add a lint that warns on #[must_use] attributes without a note

[github-actions]

Lintcheck changes for 3a164cc

Lint	Added	Removed	Changed
clippy::must_use_candidate	0	0	4657
clippy::must_use_without_note	534	0	0

This comment will be updated if you push new changes

[samueltardieu]

Note that restriction would probably be a better choice for such a lint.

[skatromb]

Thanks for the suggestion. Can you elaborate on why do you think so?
I am doubting it should be restriction, looking at the docs, they state "lints which prevent the use of Cargo features", while #[must_use] attribute could be used anywhere.

[skatromb]

Hm, but now I found another doc, which says different about this group. However, I am not sure this is applicable either

The clippy::restriction group contains lints that will restrict you from using certain parts of the Rust language. It is not recommended to enable the whole group, but rather cherry-pick lints that are useful for your code base and your use case.
Lints from this group will restrict you in some way. If you enable a restriction lint for your crate it is recommended to also fix code that this lint triggers on. However, those lints are really strict by design, and you might want to #[allow] them in some special cases, with a comment justifying that.

[samueltardieu]

This lint is restricting the use of #[must_use] unless it is accompanied with a message. This is an opiniated lint which, I think, belongs to restriction.

[rustbot]

r? @dswij

rustbot has assigned @dswij.
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

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: 7 candidates
• 7 candidates expanded to 7 candidates
• Random selection from Jarcho, dswij, llogiq, samueltardieu