fix(resolve): Warn on Self-Type Lifetime Elision for Non-Self Types

[TKanX]

View all comments

Summary

FCW lint self_lifetime_elision_not_applicable: fires when a self parameter contains & but does not syntactically contain Self.

Tracking: #140611

Context

Self-type lifetime elision in rustc_resolve relies on a name-matching hack (is_self_ty) that shallow-compares Res values between the self parameter type and the impl self type. As @compiler-errors identified in #140611, this hack cannot see through type aliases (different Res) and only fires for Struct|Union|Enum|PrimTy impls (excludes TyAlias). When it misses, find_lifetime_for_self returns Empty, self-elision silently fails, and a later parameter's lifetime may leak into the return type.

@lcnr proposed restricting self-elision to genuine Self and FCW'ing everything else. This PR implements that.

Behavior

pub struct W<'a>(&'a ());
pub type Alias<'a> = W<'a>;

Case	Code	Before	After
f1	impl Alias { fn f(self: &W) -> &() }	silent wrong '_	FCW
f2	impl Alias { fn f(self: &Alias) -> &() }	silent wrong '_	FCW
f3	impl Alias { fn f(&self) -> &() }	correct	unchanged
f4	impl W { fn f(self: &W) -> &() }	correct (hack fires)	FCW
f5	impl W { fn f(self: &Alias) -> &() }	silent wrong '_	FCW
f6	impl W { fn f(&self) -> &() }	correct	unchanged

A follow-up PR will remove the hack after a warning period.

r? @lcnr
cc @compiler-errors @traviscross

[rustbot]

Some changes occurred in compiler/rustc_hir/src/attrs

cc @jdonszelmann, @JonathanBrouwer

[lcnr]

before discussing this in a lang meeting, can you extend the PR description a bit to summarize the reasoning and discussion behind this?

[traviscross]

@rfcbot fcp concern should-example-cases-without-lifetimes

[nikomatsakis]

@rfcbot reviewed

[rustbot]

This PR modifies tests/ui/issues/. If this PR is adding new tests to tests/ui/issues/,
please refrain from doing so, and instead add it to more descriptive subdirectories.

[rustbot]

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[TKanX]

@rfcbot fcp concern should-example-cases-without-lifetimes

d1ab06e

@traviscross @scottmcm

[traviscross]

@bors try

TODO: @craterbot run mode=check-only crates=https://crater-reports.s3.amazonaws.com/pr-153692/retry-regressed-list.txt p=1

[rust-bors]

☀️ Try build successful (CI)
Build commit: 48c9d93 (48c9d93cf65d550d124c40108859fd993778c157, parent: 4b9792692fbb675174d4d2082e7c37b2bc930e71)

[traviscross]

@craterbot run mode=check-only crates=https://crater-reports.s3.amazonaws.com/pr-153692/retry-regressed-list.txt p=1

[craterbot]

👌 Experiment pr-153692-1 created and queued.
🤖 Automatically detected try build 48c9d93
🔍 You can check out the queue and this experiment's details.

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[craterbot]

🚧 Experiment pr-153692-1 is now running

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[craterbot]

🎉 Experiment pr-153692-1 is completed!
📊 151 regressed and 0 fixed (7183 total)
📊 887 spurious results on the retry-regressed-list.txt, consider a retry¹ if this is a significant amount.
📰 Open the summary report.

⚠️ If you notice any spurious failure please add them to the denylist!
ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

Footnotes

1. re-run the experiment with crates=https://crater-reports.s3.amazonaws.com/pr-153692-1/retry-regressed-list.txt ↩

[TKanX]

Can the concern be resolved now?

@scottmcm @traviscross

[traviscross]

Looking at the crater results, two other cases stand out to me:

• f(self: &W) -> &'static () — I.e., there's no elided output lifetime.
• f(self: &W) -> &() — I.e., there's only one distinct input lifetime, so this must be an f4-shape false-positive case (if it compiles otherwise).

What are your thoughts on excluding these cases as well?

[rust-bors]

☔ The latest upstream changes (presumably #157376) made this pull request unmergeable. Please resolve the merge conflicts.

[traviscross]

@rustbot author