NFC normalize lifetime identifiers#149192
Conversation
|
This is a observable language change in that previously we didn't accept struct Foo<'a>(&'a ());
fn foo<'K>(_x: Foo<'K>, _y: Foo<'K>) {}
// ^- latin K ^- kelvin KSo when we do NFC-normalize the lifetimes, we are now accepting more code (but is consistent with the original RFC and regular identifiers). This needs a lang FCP, and probably an update to the reference on NFC-normalization. @rustbot label: +I-lang-nominated |
|
I think the reference is actually fine - I'm not sure if something changed since the original bug report or the original bug report was just a generous reading of the current reference as accepting the non-normalizing implementation. The lifetime syntax page is here, and simply references the lifetime identifiers as either |
|
We talked about this on the lang call today. We agreed that idents used for lifetimes should work in the same way here as idents in other positions. @rfcbot fcp merge lang |
|
Team member @traviscross has proposed to merge this. The next step is review by the rest of the tagged team members: No concerns currently listed. Once a majority of reviewers approve (and at most 2 approvals are outstanding), this will enter its final comment period. If you spot a major issue that hasn't been raised at any point in this process, please speak up! cc @rust-lang/lang-advisors: FCP proposed for lang, please feel free to register concerns. |
|
@rfcbot reviewed |
|
🔔 This is now entering its final comment period, as per the review above. 🔔 |
|
I absolutely think we should always have lifetimes treated like tick-then-ident, so if we're doing NFC for idents then 100% agreed we should do the same for lifetimes. @rfcbot reviewed |
|
Possibly perf-sensitive, we're now calling @bors try @rust-timer queue Implementation looks fine, feel free to r=me when FCP and perf finishes. |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
NFC normalize lifetime identifiers
This comment was marked as duplicate.
This comment was marked as duplicate.
This comment has been minimized.
This comment has been minimized.
|
Finished benchmarking commit (94d5cd8): comparison URL. Overall result: ❌✅ regressions and improvements - no action neededBenchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf. @bors rollup=never Instruction countOur most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.
Max RSS (memory usage)Results (primary 1.0%)A less reliable metric. May be of interest, but not used to determine the overall result above.
CyclesResults (primary -2.8%, secondary 4.7%)A less reliable metric. May be of interest, but not used to determine the overall result above.
Binary sizeResults (primary 0.0%, secondary 0.0%)A less reliable metric. May be of interest, but not used to determine the overall result above.
Bootstrap: 467.027s -> 468.583s (0.33%) |
|
The final comment period, with a disposition to merge, as per the review above, is now complete. As the automated representative of the governance process, I would like to thank the author for their work and everyone else who contributed. |
|
@bors r+ I'm unsure if there is a measurable perf regression, or if it's just spurious. In any case, not really something we can do differently. |
|
If there is a perf regression this suggests that the normalization code might be able to be an order of magnitude faster. Which given that we're already running this on all other identifiers - and I'd guess that there are a lot more normal identifiers than lifetime ones - might lead to a fairly sizable improvement. A way to test if this is worth pursuing would be to run benchmarks with normalization commented out for normal identifiers as well. |
|
☀️ Test successful - checks-actions |
What is this?This is an experimental post-merge analysis report that shows differences in test outcomes between the merged PR and its parent PR.Comparing 430d829 (parent) -> 2cd4ee6 (this PR) Test differencesShow 6 test diffsStage 1
Stage 2
Additionally, 4 doctest diffs were found. These are ignored, as they are noisy. Job group index
Test dashboardRun cargo run --manifest-path src/ci/citool/Cargo.toml -- \
test-dashboard 2cd4ee6bcf517345dd76d7b102d56b2e873cddbc --output-dir test-dashboardAnd then open Job duration changes
How to interpret the job duration changes?Job durations can vary a lot, based on the actual runner instance |
|
Finished benchmarking commit (2cd4ee6): comparison URL. Overall result: no relevant changes - no action needed@rustbot label: -perf-regression Instruction countThis benchmark run did not return any relevant results for this metric. Max RSS (memory usage)Results (primary 1.5%, secondary -2.5%)A less reliable metric. May be of interest, but not used to determine the overall result above.
CyclesResults (secondary -0.8%)A less reliable metric. May be of interest, but not used to determine the overall result above.
Binary sizeThis benchmark run did not return any relevant results for this metric. Bootstrap: 473.205s -> 474.624s (0.30%) |
NFC normalize lifetime identifiers Fixes rust-lang#126759
Pkgsrc changes: * Update version & checksums. * Adapt patches to new vendored crates. This has so far just been verified to build on NetBSD/amd64. Upstream changes relative to 1.93.1: Version 1.94.0 (2026-03-05) ========================== Language -------- - [Impls and impl items inherit `dead_code` lint level of the corresponding traits and trait items] (rust-lang/rust#144113) - [Stabilize additional 29 RISC-V target features including large portions of the RVA22U64 / RVA23U64 profiles] (rust-lang/rust#145948) - [Add warn-by-default `unused_visibilities` lint for visibility on `const _` declarations] (rust-lang/rust#147136) - [Update to Unicode 17] (rust-lang/rust#148321) - [Avoid incorrect lifetime errors for closures] (rust-lang/rust#148329) Platform Support ---------------- - [Add `riscv64im-unknown-none-elf` as a tier 3 target] (rust-lang/rust#148790) Refer to Rust's [platform support page][platform-support-doc] for more information on Rust's tiered platform support. [platform-support-doc]: https://doc.rust-lang.org/rustc/platform-support.html Libraries --------- - [Relax `T: Ord` bound for some `BinaryHeap<T>` methods.] (rust-lang/rust#149408) Stabilized APIs --------------- - [`<[T]>::array_windows`] (https://doc.rust-lang.org/stable/std/primitive.slice.html#method.array_windows) - [`<[T]>::element_offset`] (https://doc.rust-lang.org/stable/std/primitive.slice.html#method.element_offset) - [`LazyCell::get`] (https://doc.rust-lang.org/stable/std/cell/struct.LazyCell.html#method.get) - [`LazyCell::get_mut`] (https://doc.rust-lang.org/stable/std/cell/struct.LazyCell.html#method.get_mut) - [`LazyCell::force_mut`] (https://doc.rust-lang.org/stable/std/cell/struct.LazyCell.html#method.force_mut) - [`LazyLock::get`] (https://doc.rust-lang.org/stable/std/sync/struct.LazyLock.html#method.get) - [`LazyLock::get_mut`] (https://doc.rust-lang.org/stable/std/sync/struct.LazyLock.html#method.get_mut) - [`LazyLock::force_mut`] (https://doc.rust-lang.org/stable/std/sync/struct.LazyLock.html#method.force_mut) - [`impl TryFrom<char> for usize`] (https://doc.rust-lang.org/stable/std/convert/trait.TryFrom.html#impl-TryFrom%3Cchar%3E-for-usize) - [`std::iter::Peekable::next_if_map`] (https://doc.rust-lang.org/stable/std/iter/struct.Peekable.html#method.next_if_map) - [`std::iter::Peekable::next_if_map_mut`] (https://doc.rust-lang.org/stable/std/iter/struct.Peekable.html#method.next_if_map_mut) - [x86 `avx512fp16` intrinsics] (rust-lang/rust#127213) (excluding those that depend directly on the unstable `f16` type) - [AArch64 NEON fp16 intrinsics] (rust-lang/rust#136306) (excluding those that depend directly on the unstable `f16` type) - [`f32::consts::EULER_GAMMA`] (https://doc.rust-lang.org/stable/std/f32/consts/constant.EULER_GAMMA.html) - [`f64::consts::EULER_GAMMA`] (https://doc.rust-lang.org/stable/std/f64/consts/constant.EULER_GAMMA.html) - [`f32::consts::GOLDEN_RATIO`] (https://doc.rust-lang.org/stable/std/f32/consts/constant.GOLDEN_RATIO.html) - [`f64::consts::GOLDEN_RATIO`] (https://doc.rust-lang.org/stable/std/f64/consts/constant.GOLDEN_RATIO.html) These previously stable APIs are now stable in const contexts: - [`f32::mul_add`] (https://doc.rust-lang.org/stable/std/primitive.f32.html#method.mul_add) - [`f64::mul_add`] (https://doc.rust-lang.org/stable/std/primitive.f64.html#method.mul_add) Cargo ----- - Stabilize the config include key. The top-level include config key allows loading additional config files, enabling better organization, sharing, and management of Cargo configurations across projects and environments. [docs] (https://doc.rust-lang.org/nightly/cargo/reference/config.html#including-extra-configuration-files) [#16284] (rust-lang/cargo#16284) - Stabilize the pubtime field in registry index. This records when a crate version was published and enables time-based dependency resolution in the future. Note that crates.io will gradually backfill existing packages when a new version is published. Not all crates have pubtime yet. [#16369] (rust-lang/cargo#16369) [#16372] (rust-lang/cargo#16372) - Cargo now parses [TOML v1.1](https://toml.io/en/v1.1.0) for manifests and configuration files. Note that using these features in Cargo.toml will raise your development MSRV, but the published manifest remains compatible with older parsers. [#16415] (rust-lang/cargo#16415) - [Make `CARGO_BIN_EXE_<crate>` available at runtime ] (rust-lang/cargo#16421) Compatibility Notes ------------------- - [Forbid freely casting lifetime bounds of `dyn`-types] (rust-lang/rust#136776) - [Make closure capturing have consistent and correct behaviour around patterns] (rust-lang/rust#138961) Some finer details of how precise closure captures get affected by pattern matching have been changed. In some cases, this can cause a non-move closure that was previously capturing an entire variable by move, to now capture only part of that variable by move, and other parts by borrow. This can cause the borrow checker to complain where it previously didn't, or cause `Drop` to run at a different point in time. - [Standard library macros are now imported via prelude, not via injected ` #[macro_use]`] (rust-lang/rust#139493) This will raise an error if macros of the same name are glob imported. For example if a crate defines their own `matches` macro and then glob imports that, it's now ambiguous whether the custom or standard library `matches` is meant and an explicit import of the name is required to resolve the ambiguity. One exception is `core::panic` and `std::panic`, if their import is ambiguous a new warning ([`ambiguous_panic_imports`] (rust-lang/rust#147319)) is raised. This may raise a new warning ([`ambiguous_panic_imports`] (rust-lang/rust#147319)) on `#![no_std]` code glob importing the std crate. Both `core::panic!` and `std::panic!` are then in scope and which is used is ambiguous. - [Don't strip shebang in expression-context `include!(…)`s] (rust-lang/rust#146377) This can cause previously working includes to no longer compile if they included files which started with a shebang. - [Ambiguous glob reexports are now also visible cross-crate] (rust-lang/rust#147984) This unifies behavior between local and cross-crate errors on these exports, which may introduce new ambiguity errors. - [Don't normalize where-clauses before checking well-formedness] (rust-lang/rust#148477) - [Introduce a future compatibility warning on codegen attributes on body-free trait methods] (rust-lang/rust#148756) These attributes currently have no effect in this position. - [On Windows `std::time::SystemTime::checked_sub_duration` will return `None` for times before the Windows epoch (1/1/1601)] (rust-lang/rust#148825) - [Lifetime identifiers such as `'a` are now NFC normalized] (rust-lang/rust#149192). - [Overhaul filename handling for cross-compiler consistency] (rust-lang/rust#149709) Any paths emitted by compiler now always respect the relative-ness of the paths and `--remap-path-prefix` given originally. One side-effect of this change is that paths emitted for local crates in Cargo (path dependencies and workspace members) are no longer absolute but relative when emitted as part of a diagnostic in a downstream crate. Internal Changes ---------------- These changes do not affect any public interfaces of Rust, but they represent significant improvements to the performance or internals of rustc and related tools. - [Switch to `annotate-snippets` for error emission] (rust-lang/rust#150032) This should preserve mostly the same outputs in rustc error messages.
Fixes #126759