Don't warn for missing_doc_examples when item is #[doc(hidden)]

[GuillaumeGomez]

r? @jyn514

[jyn514]

I have a secondary concern too: it seems like a lot of code between this and missing_docs is being duplicated. Is there anyway to make them more consistent with one another? It seems a shame to have to update the same logic in two places.

[jyn514]

@Dylan-DPC why did you mark this as waiting-on-team?

[GuillaumeGomez]

I am the team maybe? XD

[Dylan-DPC-zz]

This is pending on discussion between both of you and anyone else from the team :P

[jyn514]

This is pending on discussion between both of you and anyone else from the team :P

Not that I'm aware of.

[bors]

☔ The latest upstream changes (presumably #82756) made this pull request unmergeable. Please resolve the merge conflicts.

[jyn514]

This makes missing_doc_examples inconsistent with missing_docs:

warning: missing documentation for the crate
 --> hidden.rs:1:1
  |
1 | / #![warn(missing_docs)]
2 | | #[doc(hidden)]
3 | | pub fn f() {}
  | |_____________^
  |
note: the lint level is defined here
 --> hidden.rs:1:9
  |
1 | #![warn(missing_docs)]
  |         ^^^^^^^^^^^^

I do think the new behavior is better, but it also seems strange to have the two behave differently ... maybe we should also update missing_docs to have this behavior? But that requires keeping track of doc(hidden) in rustc itself, which would be a pretty big change from how it works now.

@Manishearth @ollie27 what do you think?

[jyn514]

Oh wait I'm just dumb, that was warning that the crate root is missing docs, not that the item is missing docs. Adding crate-level docs fixes the issue:

$ cat hidden.rs 
//! some docs
#![warn(missing_docs)]
#[doc(hidden)]
pub fn f() {}

@GuillaumeGomez r=me with the merge conflict fixed

[GuillaumeGomez]

@bors: r=jyn514

[bors]

📌 Commit 1683cb1 has been approved by jyn514