[NUAAwqy]

What does this PR try to resolve?

Fixes #17609

For dependencies defined with table header syntax (e.g. [dependencies.foo]), the value span returned by get_key_value covers the whole table header, including its closing ]. The unused-dependency diagnostic underlined key.start..value.end, so the underline dragged the stray ] into the diagnostic:

warning: unused dependency `sp1-zkvm`
  --> Cargo.toml:14:15
   |
14 | [dependencies.sp1-zkvm]
   |               ^^^^^^^^^

With this change, header-defined tables underline only the dependency key:

warning: unused dependency `sp1-zkvm`
  --> Cargo.toml:14:15
   |
14 | [dependencies.sp1-zkvm]
   |               ^^^^^^^^

Inline tables (foo = { .. }) and plain key-value entries (foo = "1.0") are unaffected and keep the full key..value span. The same fix is applied to both the normal-dependency and build-dependency call sites in unused_dependencies.rs.

Notes to reviewers

The probe that motivated the detection rule: for a header-defined table the value span starts with [ (it covers the entire header), while for an inline table it starts with {. dep_underline_span keys off that distinction.

Verified locally:

• cargo test --test testsuite lints::unused_dependencies — all 25 tests pass, including the new unused_dep_table_header_syntax.
• The new test fails on master (underline includes ]) and passes with the fix.

---

🤖 LLM involvement: An LLM assistant (Claude) was used to investigate the issue, draft the fix and the regression test, and prepare this PR. All code was reviewed and the tests were run locally by the author.

[rustbot]

Thanks for the pull request, and welcome! The Rust Project has assigned @epage (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.

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: @Muscraft, @epage, @weihanglo
• @Muscraft, @epage, @weihanglo expanded to Muscraft, epage, weihanglo
• Random selection from Muscraft, epage, weihanglo

[weihanglo]

This PR, as disclosed, was created with LLM assistance, but has no pre-arranged reviewers for LLM generated changes, so closing.

[weihanglo]

Also your PR activity seems to be LLM automation. This is not allowed in any repositories under rust-lang organization per our policies. We are thus restricting you from creating pull requests and comments.

Thanks for your understanding.

Weihang in the name of the mod team