Transition future compat lints to {ERROR, DENY} - Take 2

[Centril]

Follow up to #63247 implementing #63247 (comment).

• legacy_ctor_visibility (ERROR) -- closes Tracking issue for legacy_constructor_visibility compatibility lint #39207
• legacy_directory_ownership (ERROR) -- closes legacy_directory_ownership future-compatibility warnings #37872
• safe_extern_static (ERROR) -- closes Tracking issue for safe_extern_statics compatibility lint #36247
• parenthesized_params_in_types_and_modules (ERROR) -- closes parenthesized_params_in_types_and_modules future-compatibility warnings #42238
• duplicate_macro_exports (ERROR)
• nested_impl_trait (ERROR) -- closes Tracking issue for nested_impl_trait compatibility lint #59014
• ill_formed_attribute_input (DENY) -- transitions Tracking issue for future-incompatibility lint ill_formed_attribute_input #57571
• patterns_in_fns_without_body (DENY) -- transitions Tracking issue for future-incompatibility lint patterns_in_fns_without_body #35203

r? @varkor
cc @petrochenkov

[Centril]

The PR builder is finally happy. 🎉

[petrochenkov]

Also, legacy_ctor_visibility had a large number of regressions in the crater run in #63247, but most of them from unstable syntex_syntax (expected function, found struct ast::Name), so it should be ok.

[Centril]

@petrochenkov @varkor This should be ready again for review now.

[oli-obk]

I'm all for this, but the precedent also had fix-PRs open for the crates that will break and did a reverse dependency analysis.

I know that doing so is a lot of work, but we do have our stability story. Picking an arbitrary threshold (like 5) of broken crates where it's ok to go to an error, seems a bit odd to me. Then again, we have to do this PR at some point and the forward compat lints have been triggering for a while.

I'm unsure how to proceed, thus my question about talking this through in a team meeting or just giving everyone a chance to sign off or comment.

[oli-obk]

Seems like all the crates affected are abandoned (which would explain why they haven't been fixed).

@bors r+

[bors]

📌 Commit 574d2b8 has been approved by oli-obk

[bors]

⌛ Testing commit 574d2b8 with merge 230ebbbe9d0ea2faf5fe586b8f3f049b6f571c4b...

[Centril]

@bors retry rolled up.

[bors]

⌛ Testing commit 574d2b8 with merge 1465351c3dbabfa79ff87afa4e90c34d77be68ae...

[rust-highfive]

Your PR failed (pretty log, raw log). Through arcane magic we have determined that the following fragments from the build log may contain information about the problem.

Click to expand the log.

I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact @TimNN. (Feature Requests)