possible misuse of #[zenoh_macros::unstable] and #[zenoh_macros::unstable_doc]

[milyin]

Describe the release item

The #[zenoh_macros::unstable] is fully equal to #[cfg(attribute="unstable")]
The #[zenoh_macros::unstable_doc] gives additional functionality: it adds information to documnatation that this is unstable feature.
Also just #[cfg(attribute="unstable")] is actively used

The proposal is:

• remove #[zenoh_macros::unstable]: replace it to #[cfg(attribute="unstable")] for pieces of code and to #[zenoh_macros::unstable_doc] where the documentation should be corrected.
• add check to #[zenoh_macros::unstable_doc] forbidding it's usage if no documentation for item exists

[wyfo]

The #[zenoh_macros::unstable] is fully equal to #[cfg(attribute="unstable")]

I'm not sure to following you. #[zenoh_macros::unstable] is equivalent to #[cfg(attribute="unstable")]#[zenoh_macros::unstable_doc], not just #[cfg(attribute="unstable")].

If I've understood correctly, zenoh_macros::unstable_doc is for items of "unstable" crates, like zenoh-shm, which are reexported in zenoh. These crate don't have a unstable feature, so zenoh_macros::unstable would not have sense.
So both are needed.

add check to #[zenoh_macros::unstable_doc] forbidding it's usage if no documentation for item exists

I'm +1 on that.

[milyin]

About equivalence - it was my mistake, zenoh_macros::unstable also updates documentation.
Yes, I found that we can't remove unstable_doc - it seems if was specially created for adding "unstable" documentation comment to crate without scattering unstable feature everywhere.
About forbidding usage of macro when no documentation - seems like it adds more problems than solves. It becomes easy to add documentation and forget to add zenoh_macros::unstable.
So closing it as not a bug.