[johncarmack1984]

#[deprecated] on a use item is accepted and then ignored: deprecation is checked on the resolved item, not on the imports a path went through, so deprecating a re-export that way warns nobody (#30827). This adds ineffective_deprecated_reexports, a warn-by-default lint that says so at the attribute, notes that it's a long-standing bug, and suggests what does work: deprecating the item itself, or a #[deprecated] type alias or wrapper function where one applies.

Choices:

• A new warn-by-default lint rather than useless_deprecated, which is deny-by-default. The review on Add lint rule for #[deprecated] on re-exports (rebase) #142731 asked for a lint instead of an error.
• Emitted from the stability checker in rustc_passes, next to ineffective_unstable_reexports (lint ineffective #[unstable] annotations on re-exports #161178).
• The help is prose, not a structured suggestion, because no single rewrite fits: a type alias doesn't carry tuple or unit struct constructors or use Alias::Variant, a function needs a wrapper, and a module or macro re-export has no equivalent.
• core (alloc::LayoutErr) and std (collections::Bound) each carry one of these attributes. Both get an item-level #[expect], which std's deny-warnings build turns into an error once the lint stops firing there; core gets a FIXME, std already had one. The run-rustfix test in tests/ui/imports carries its expectation on the item: a crate-level one is unfulfilled in the fixed output, which deletes the annotated use.

The fix for #30827 should delete this lint.

Supersedes #142731 (I asked GrigorenkoPV there on Oct 22 whether they planned to return to it and haven't heard back). Closes #142436.

r? @WaffleLapkin

LLM disclosure: I wrote this lint. Claude Code (Fable 5.1, then Opus 5.5, both max effort) evaluated the design options in the issue's history, drafted the implementation, docs and tests, and ran the test suites. Codex reviewed. I pared and reviewed every line.

[rustbot]

WaffleLapkin is currently at their maximum review capacity.
They may take a while to respond.

[rustbot]

Thanks for the pull request, and welcome! The Rust Project has assigned @WaffleLapkin (or someone else) to review your changes, you should hear from them (or someone else) within the next two weeks.

Please see the contribution instructions and our LLM policy for more information.

[theemathas]

What exactly is the LLM involvement here? What does "I wrote this lint" entail?

[Kivooeo]

docs

Hi, first of all, all user-facing components including PR description should be written manually from start to finish; otherwise, this violates our policy. Second, you should have found a pre-arranged LLM reviewer in #llm-reviews before opening a PR.

More you can read in our LLM policy.