Add a new `mismatched-lifetime-syntaxes` lint #138677

shepmaster · 2025-03-18T17:58:06Z

The lang-team discussed this and I attempted to summarize their decision. The summary-of-the-summary is:

Using two different kinds of syntax for elided lifetimes is confusing. In rare cases, it may even lead to unsound code! Some examples:
```
// Lint will warn about these
fn(v: ContainsLifetime) -> ContainsLifetime<'_>;
fn(&'static u8) -> &u8;
```
Matching up references with no lifetime syntax, references with anonymous lifetime syntax, and paths with anonymous lifetime syntax is an exception to the simplest possible rule:
```
// Lint will not warn about these
fn(&u8) -> &'_ u8;
fn(&'_ u8) -> &u8;
fn(&u8) -> ContainsLifetime<'_>;
```
Having a lint for consistent syntax of elided lifetimes will make the future goal of warning-by-default for paths participating in elision much simpler.

This new lint attempts to accomplish the goal of enforcing consistent syntax. In the process, it supersedes and replaces the existing elided-named-lifetimes lint, which means it starts out life as warn-by-default.

rustbot · 2025-03-18T17:58:11Z

r? @wesleywiser

rustbot has assigned @wesleywiser.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

compiler/rustc_lint/src/lints.rs

shepmaster · 2025-03-18T18:26:44Z

Thanks to:

@GrigorenkoPV for the original implementation of elided_named_lifetimes.
@compiler-errors for the push to write the lint using HIR instead of in rustc_resolve.
@estebank for the diagnostic help and wording feedback.

tests/ui/lifetimes/mismatched-elided-lifetime-styles.rs

compiler/rustc_lint/src/lifetime_style.rs

Addressed.

bors · 2025-03-22T11:05:50Z

☔ The latest upstream changes (presumably #138719) made this pull request unmergeable. Please resolve the merge conflicts.

rustbot · 2025-03-23T00:48:30Z

Some changes occurred to the CTFE machinery

cc @RalfJung, @oli-obk, @lcnr

rustc_macros::diagnostics was changed

cc @davidtwco, @compiler-errors, @TaKO8Ki

Some changes occurred in exhaustiveness checking

cc @Nadrieril

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

rustbot · 2025-03-24T17:29:18Z

Some changes occurred in src/tools/clippy

cc @rust-lang/clippy

bors · 2025-03-28T07:30:13Z

☔ The latest upstream changes (presumably #138965) made this pull request unmergeable. Please resolve the merge conflicts.

bors · 2025-04-04T02:32:58Z

☔ The latest upstream changes (presumably #139336) made this pull request unmergeable. Please resolve the merge conflicts.

bors · 2025-04-15T08:55:32Z

☔ The latest upstream changes (presumably #139826) made this pull request unmergeable. Please resolve the merge conflicts.

bors · 2025-04-17T05:43:08Z

☔ The latest upstream changes (presumably #139938) made this pull request unmergeable. Please resolve the merge conflicts.

shepmaster · 2025-04-24T13:46:37Z

r? compiler

bors · 2025-05-02T18:20:51Z

☔ The latest upstream changes (presumably #140581) made this pull request unmergeable. Please resolve the merge conflicts.

traviscross

@rustbot author

tests/ui/suggestions/lifetimes/missing-lifetimes-in-signature.rs

tests/ui/object-lifetime/object-lifetime-default-elision.rs

traviscross · 2025-05-03T08:51:44Z

tests/ui/lifetimes/mismatched-lifetime-syntaxes.rs

+
+struct S(u8);
+
+fn named_ref_to_hidden_ref<'a>(v: &'a u8) -> &u8 {


This is named to elided.

I do not agree that this test name should be changed.

These test names are based on the syntax used, not the behavior. If they were not, there would be no naming scheme to disambiguate these materially different cases:

fn named_ref_to_elided_ref<'a>(v: &'a u8) -> &u8 fn named_ref_to_elided_ref<'a>(v: &'a u8) -> &'_ u8

That enum was added recently by you as part of this same work. I'd probably recommend changing that as well. Otherwise it's going to be one of those continually confusing areas where the compiler and the lang team / Reference disagree about terminology.

The lang team reviews UI tests to evaluate stabilizations and for other such purposes. If we want to actually use the terms differently between how we've decided to use them on the lang side and how we use them in the compiler, I feel like we're going to need to write this down carefully somewhere and then cite it widely in all the places we divergently use these terms.

(I'd prefer we find a way to avoid that and to either use the same terms to mean the same things or, if we can't align our concepts, at least to use non-overlapping terms.)

This thread consolidates:

Add a new mismatched-lifetime-syntaxes lint #138677 (comment)

Add a new mismatched-lifetime-syntaxes lint #138677 (comment)

That enum was added recently by you as part of this same work

Yes. The only reason it's not part of this PR is that it was blocking other compiler contributors and was extracted to a separate PR. I link to it as a listing of the choices available, not to prove its authoritativeness.

one of those continually confusing areas where the compiler and the lang team / Reference disagree about terminology

I see the point and wish there were not confusion, but there are two related yet different concepts:

What the user typed / what the source code shows (what I'm calling "syntax")

The higher-level meaning (what I'm calling "style")

I assert that using the same name for the syntax of &T and &'_ T will likely be confusing to users ¹ and will almost certainly be confusing to compiler developers. That's why I've chosen the implementation path of having the two enums (LifetimeSyntax, LifetimeSource) as the primary representation through the compiler and restricted the confusing meaning to the lint.

User-facing parts of the lint (the documentation, lint messages) should indeed use the lang-team-decided terminology of named/elided/hidden where needed.

For my own part (which I don't think I've done a good job expressing), I think that saying that fn x(&u8) -> ContainsLifetime has one "elided lifetime" and one "hidden lifetime" when the action of lifetime elision occurs equally for both will already confuse users ².

I'd prefer we find a way to avoid that and to either use the same terms to mean the same things or, if we can't align our concepts, at least to use non-overlapping terms.

That would be wonderful, but I'm not seeing how to align the terms completely because of the intrinsic differences. The source code allows for the product of {Reference, Path} ✕ {NoSyntax, AnonymousLifetime, NamedLifetime} but meaning-wise we group it as

(Reference, NoSyntax) / (Reference, AnonymousLifetime) / (Path, AnonymousLifetime) ("elided")

(Path, NoSyntax) ("hidden")

(Reference, NamedLifetime) / (Path, NamedLifetime) ("named")

At least until the HIR level, I argue that the compiler should correspond to what the user entered.

Footnotes

e.g. "oh, change this to use an 'elided lifetime'"... "what kind?" ... "i dunno what to call it" ↩

e.g. "so since this one is hidden and this one is elided, that means that lifetime elision doesn't occur here?" ... "well, no, elision does occur, it's just not called elided" ↩

If the concern is syntactic, and you want to apply the same categorization to both references and paths, I'd suggest that either can have an "explicit" lifetime parameter (e.g. &'a (), &'_ (), Ref<'a>, Ref<'_>) or an "implicit" one (e.g. &(), Ref), and that explicit lifetime parameters can be either a "bound" lifetime (e.g. &'a (), Ref<'a>), the "anonymous" one (e.g. &'_ (), Ref<'_>), or 'static.

tests/ui/lifetimes/mismatched-lifetime-syntaxes.rs

traviscross · 2025-05-03T08:58:30Z

tests/ui/lifetimes/mismatched-lifetime-syntaxes.rs

+        ContainsLifetime(v)
+    }
+
+    fn hidden_path_to_hidden_ref(v: ContainsLifetime) -> &u8 {


This is a mismatch and should fire the lint.

tests/ui/lifetimes/mismatched-lifetime-syntaxes.rs

traviscross · 2025-05-03T09:01:11Z

tests/ui/lifetimes/mismatched-lifetime-syntaxes.rs

+        v
+    }
+
+    fn hidden_ref_to_hidden_path(v: &u8) -> ContainsLifetime {


This is a mismatch and should fire the lint.

tests/ui/lifetimes/lifetime-errors/ex1-return-one-existing-name-if-else-using-impl-3.rs

rustbot · 2025-05-03T09:17:28Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

shepmaster

I haven't been able to address the corrections I agree with, but wanted to comment quickly in an attempt to reduce the month-long review cycle by a few hours.

tests/ui/suggestions/lifetimes/missing-lifetimes-in-signature.rs

tests/ui/object-lifetime/object-lifetime-default-elision.rs

tests/ui/lifetimes/lifetime-errors/ex1-return-one-existing-name-if-else-using-impl-3.rs

shepmaster · 2025-05-03T12:04:06Z

tests/ui/lifetimes/mismatched-lifetime-syntaxes.rs

+
+struct S(u8);
+
+fn named_ref_to_hidden_ref<'a>(v: &'a u8) -> &u8 {


I do not agree that this test name should be changed.

These test names are based on the syntax used, not the behavior. If they were not, there would be no naming scheme to disambiguate these materially different cases:

fn named_ref_to_elided_ref<'a>(v: &'a u8) -> &u8 fn named_ref_to_elided_ref<'a>(v: &'a u8) -> &'_ u8

tests/ui/lifetimes/mismatched-lifetime-syntaxes.rs

shepmaster · 2025-05-03T13:07:27Z

tests/ui/lifetimes/mismatched-lifetime-syntaxes.rs

+        ContainsLifetime(v)
+    }
+
+    fn hidden_path_to_hidden_ref(v: ContainsLifetime) -> &u8 {


shepmaster · 2025-05-03T13:07:32Z

tests/ui/lifetimes/mismatched-lifetime-syntaxes.rs

+        v
+    }
+
+    fn hidden_ref_to_hidden_path(v: &u8) -> ContainsLifetime {


tests/ui/lifetimes/mismatched-lifetime-syntaxes.rs

rustbot assigned wesleywiser Mar 18, 2025

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. T-libs Relevant to the library team, which will review and decide on the PR/issue. labels Mar 18, 2025

shepmaster mentioned this pull request Mar 18, 2025

Split elided_lifetime_in_paths into tied and untied #120808

Draft

3 tasks

This comment has been minimized.

Sign in to view

shepmaster commented Mar 18, 2025

View reviewed changes

compiler/rustc_lint/src/lints.rs Outdated Show resolved Hide resolved

shepmaster force-pushed the consistent-elided-lifetime-syntax branch from 478a98e to e15c083 Compare March 18, 2025 21:01

This comment has been minimized.

Sign in to view

traviscross reviewed Mar 19, 2025

View reviewed changes

tests/ui/lifetimes/mismatched-elided-lifetime-styles.rs Outdated Show resolved Hide resolved

traviscross previously requested changes Mar 19, 2025

View reviewed changes

compiler/rustc_lint/src/lifetime_style.rs Outdated Show resolved Hide resolved

shepmaster force-pushed the consistent-elided-lifetime-syntax branch from e15c083 to ca86221 Compare March 20, 2025 13:34

shepmaster changed the title ~~Add a new mismatched_elided_lifetime_styles lint~~ Add a new mismatched-lifetime-syntaxes lint Mar 20, 2025

This comment has been minimized.

Sign in to view

shepmaster force-pushed the consistent-elided-lifetime-syntax branch from ca86221 to 6b4ba02 Compare March 20, 2025 13:57

This comment has been minimized.

Sign in to view

shepmaster force-pushed the consistent-elided-lifetime-syntax branch from 6b4ba02 to 7ff70d4 Compare March 23, 2025 00:48

rustbot added the A-translation Area: Translation infrastructure, and migrating existing diagnostics to SessionDiagnostic label Mar 23, 2025

This comment has been minimized.

Sign in to view

shepmaster force-pushed the consistent-elided-lifetime-syntax branch from 7ff70d4 to 3a95f66 Compare March 24, 2025 17:29

This comment has been minimized.

Sign in to view

shepmaster force-pushed the consistent-elided-lifetime-syntax branch from 3a95f66 to 1625496 Compare March 24, 2025 20:01

This comment has been minimized.

Sign in to view

shepmaster force-pushed the consistent-elided-lifetime-syntax branch from 1625496 to 1cd8999 Compare March 25, 2025 00:55

shepmaster mentioned this pull request Mar 28, 2025

Remove kw::Empty uses from hir::Lifetime::ident #138965

Merged

shepmaster force-pushed the consistent-elided-lifetime-syntax branch 2 times, most recently from 85b9c5a to 573f332 Compare March 31, 2025 20:22

This comment has been minimized.

Sign in to view

shepmaster force-pushed the consistent-elided-lifetime-syntax branch from 573f332 to 702e59d Compare April 1, 2025 13:39

shepmaster force-pushed the consistent-elided-lifetime-syntax branch 2 times, most recently from e5616d0 to e6dafb6 Compare April 7, 2025 16:44

shepmaster force-pushed the consistent-elided-lifetime-syntax branch from e6dafb6 to 5611e54 Compare April 15, 2025 12:34

shepmaster force-pushed the consistent-elided-lifetime-syntax branch from 5611e54 to 8d56830 Compare April 24, 2025 13:44

rustbot assigned BoxyUwU and unassigned wesleywiser Apr 24, 2025

shepmaster added 2 commits May 2, 2025 15:36

Introduce the mismatched_lifetime_syntaxes lint

ce3327c

Replace elided_named_lifetimes with mismatched_lifetime_syntaxes

9da1ebb

shepmaster force-pushed the consistent-elided-lifetime-syntax branch from 8d56830 to 9da1ebb Compare May 2, 2025 19:57

This was referenced May 3, 2025

lifetime elision for return type can refer to an explicitly named region #121152

Open

Oddity with lifetime elision and type aliases #140611

Open

traviscross reviewed May 3, 2025

View reviewed changes

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels May 3, 2025

shepmaster commented May 3, 2025

View reviewed changes


		struct S(u8);

		fn named_ref_to_hidden_ref<'a>(v: &'a u8) -> &u8 {

Add a new mismatched-lifetime-syntaxes lint #138677

Are you sure you want to change the base?

Add a new mismatched-lifetime-syntaxes lint #138677

Conversation

shepmaster commented Mar 18, 2025 • edited Loading

rustbot commented Mar 18, 2025

This comment has been minimized.

shepmaster commented Mar 18, 2025

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

bors commented Mar 22, 2025

rustbot commented Mar 23, 2025

This comment has been minimized.

rustbot commented Mar 24, 2025

This comment has been minimized.

This comment has been minimized.

bors commented Mar 28, 2025

This comment has been minimized.

bors commented Apr 4, 2025

bors commented Apr 15, 2025

bors commented Apr 17, 2025

shepmaster commented Apr 24, 2025

bors commented May 2, 2025

traviscross left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

traviscross May 3, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Footnotes

traviscross May 3, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rustbot commented May 3, 2025

shepmaster left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Add a new `mismatched-lifetime-syntaxes` lint #138677

Add a new `mismatched-lifetime-syntaxes` lint #138677

shepmaster commented Mar 18, 2025 •

edited

Loading

traviscross May 3, 2025 •

edited

Loading

traviscross May 3, 2025 •

edited

Loading