Make unknown/renamed/removed lints passed via command line respect lint levels

[weihanglo]

Proposal

Background

When you specify an unknown/renamed/removed lint name in Rust code, for example #![deny(foo)], the compiler respects the lint system and emits a warning by default^1.

However, if it is from the command line, such as -D foo, the compiler simply emits an error and exits^2. There is no way to recover from the error, since those errors are emitted directly via Session, not under the control of LintContext or LintLevelsbuilder.

Possible solution

This MCP proposes that those errors should be considered as normal lints and emitted by the lint system. Specifically,

• Replace CheckNameRenamed error with renamed-and-removed-lints.
• Replace CheckNameRemoved error with renamed-and-removed-lints.
• Replace CheckNameDeprecated error with renamed-and-removed-lints.
• Replace CheckNameUnknown error with unknown-lints

The proposed behavior is more consistent from both source-code level and end-user perspective.

User-facing impact

Compiler invocations that rely on the aforementioned always-error-out behavior will become valid. This is easy to fix by adding -D unknown-lints or -D renamed-and-removed-lints as a command line argument (tbh not as easy as it looks like[^3]).

Motivation

With -Zlints unstable feature in Cargo, you can specify lint rules and levels in Cargo.toml. Under the hood Cargo aggregates those settings and passes via command line to the compiler.

Without this MCP, it is not backward compatible when a user adds a new lint, and then use an old version of the compiler. The compiler will refuse to compile even unknown-lints lint is warn by default not deny.

See https://github.com/rust-lang/cargo/issues/12495.

Compiler version

$ rustc +nightly -vV
rustc 1.74.0-nightly (58eefc33a 2023-08-24)
binary: rustc
commit-hash: 58eefc33adf769a1abe12ad94b3e6811185b4ce5
commit-date: 2023-08-24
host: aarch64-apple-darwin
release: 1.74.0-nightly
LLVM version: 17.0.0

[^3]: This is another command line argument precedence issue. Think about this. Which should emit an unknown lint error? -Dunknown-lints -Dfoo or -Dfoo -Dunknown-lints? Or should we make these lints a special priority like warnings lint group? That may lead to https://github.com/rust-lang/rust/issues/75668 though. Anyway, this belongs to a separate MCP.

Mentors or Reviewers

Not sure. Perhaps @estebank or @compiler-errors?

Process

The main points of the Major Change Process are as follows:

• [x] File an issue describing the proposal.
• [ ] A compiler team member or contributor who is knowledgeable in the area can second by writing @rustbot second.
  • Finding a "second" suffices for internal changes. If however, you are proposing a new public-facing feature, such as a -C flag, then full team check-off is required.
  • Compiler team members can initiate a check-off via @rfcbot fcp merge on either the MCP or the PR.
• [ ] Once an MCP is seconded, the Final Comment Period begins. If no objections are raised after 10 days, the MCP is considered approved.

You can read more about Major Change Proposals on forge.

Comments

This issue is not meant to be used for technical discussion. There is a Zulip stream for that. Use this issue to leave procedural comments, such as volunteering to review, indicating that you second the proposal (or third, etc), or raising a concern that you would like to be addressed.

[rustbot]

This issue is not meant to be used for technical discussion. There is a Zulip stream for that. Use this issue to leave procedural comments, such as volunteering to review, indicating that you second the proposal (or third, etc), or raising a concern that you would like to be addressed.

cc @rust-lang/compiler @rust-lang/compiler-contributors

[davidtwco]

@rustbot second