Sized Hierarchy #137944

davidtwco · 2025-03-03T14:51:33Z

This patch implements rust-lang/rfcs#3729. It introduces two new traits to the standard library, MetaSized and PointeeSized, and makes MetaSized and Sized into const traits (relying on unstable feature(const_trait_impl)). See the RFC for the rationale behind these traits and to discuss whether this change makes sense in the abstract.

These traits are unstable (as is their constness), so users cannot refer to them without opting-in to feature(sized_hierarchy). These traits are not behind cfgs as this would make implementation unfeasible, there would simply be too many cfgs required to add the necessary bounds everywhere. So, like Sized, these traits are automatically implemented by the compiler.

RFC 3729 describes migrations which are necessary to preserve backwards compatibility given the introduction of these traits, which are implemented and as follows:

On the current edition, Sized is rewritten as const Sized
- If the sized_hierarchy feature is enabled, then an edition migration lint to rewrite the bound to const Sized will be emitted.
- On the next edition, non-const Sized will resume being the default bound.
On the current edition, ?Sized is rewritten as const MetaSized
- If the sized_hierarchy feature is enabled, then an edition migration lint to rewrite the bound to const MetaSized will be emitted.
- On the next edition, writing ?Sized will be prohibited.
On the current edition, const MetaSized is added as a default supertrait for all traits w/out an explicit sizedness supertrait already.
- If the sized_hierarchy feature is enabled, then an edition migration lint to add an explicit const MetaSized supertrait will be emitted.
- On the next edition, there is no default const MetaSized supertrait.

Each of these migrations is not conditional on whether the item being migrated needs the migration to the stricter bound - this would be preferable but is not yet implemented (if it is possible to implement). All diagnostic output should remain the same (showing ?Sized even if the compiler sees const MetaSized) unless the sized_hierarchy feature is enabled.

Due to the use of unstable extern types in the standard library and rustc, some bounds in both projects have had to be relaxed already - this is unfortunate but unavoidable so that these extern types can continue to be used where they were before. Performing these relaxations in the standard library and rustc are desirable longer-term anyway, but some bounds are not as relaxed as they ideally would be due to the inability to relax Deref::Target (this will be investigated separately).

It is hoped that this is implemented such that it could be merged and these traits could exist "under the hood" without that being observable to the user (other than in any performance impact this has on the compiler, etc). Only once sized_hierarchy is stabilised would edition migration lints start to be emitted and diagnostic output show the "real" sizedness traits behind-the-scenes, rather than ?Sized. Some details might leak through due to the standard library relaxations, but this has not been observed in test output.

Notes:

Any commits starting with "upstream:" can be ignored, as these correspond to other upstream PRs that this is based on which have yet to be merged.
This best reviewed commit-by-commit. I've attempted to make the implementation easy to follow and keep similar changes and test output updates together.
- Each commit has a short description describing its purpose.
- This patch is large but it's primarily in the test suite (library: +573/-184, compiler: +1268/-310, tests: +3720/-452).
It is expected that this will have performance regressions initially and I'll aim to resolve those prior to merging if possible.
- I'd appreciate feedback on how best to go about this from those familiar with the type system.
On my local machine, this passes all of the test suites, a stage two build and a tidy check.
PointeeSized is a different name from the RFC just to make it more obvious that it is different from std::ptr::Pointee but all the names are yet to be bikeshed anyway.

Fixes #79409.

r? @ghost (I'll discuss this with relevant teams to find a reviewer)

davidtwco · 2025-03-03T15:21:57Z

~~I can reproduce this locally but I have no idea why it would be related to this patch.~~ Clippy needed adjusting.

compiler/rustc_type_ir/src/flags.rs

compiler/rustc_passes/src/check_attr.rs

tests/ui/target-feature/tied-features-no-implication-1.rs

traviscross · 2025-03-04T01:46:50Z

cc @rust-lang/lang

tmiasko · 2025-03-04T10:12:47Z

Does this perhaps fix #127336 by rejecting it?

davidtwco · 2025-03-06T10:22:24Z

Does this perhaps fix #127336 by rejecting it?

It doesn't currently.

These tests just need blessing, they don't have any interesting behaviour changes. Some of these tests have new errors because `LegacyReceiver` cannot be proven to be implemented now that it is also testing for `MetaSized` - but this is just a consequence of the other errors in the test.

While there are no non-`const Sized` types in the language, this testing attribute will make it possible to force this for testing purposes.

Makes `{Meta,}Sized` into const traits and implements them in the trait solvers so everything is `Sized` except ADTs with the `rustc_non_const_sized` attribute.

Now that `MetaSized` and `Sized` are const traits, this must be reflected in minicore.

Now that `MetaSized` and `Sized` are const traits, this must be reflected in tests without minicore which define these language items.

With const sizedness traits, the correct migration is from `Sized` to `const Sized` and from `?Sized` to `const MetaSized`.

There are some uses of `min_specialization` in the standard library which now see `HostEffectPredicate` - this shouldn't prevent specialization for the sizedness marker traits that are now const.

Some of the host effect predicates end up with `TraitGoalProvenVia::ParamEnv` combined with `CandidateSource::BuiltinImpl` and are discarded as a consequence. This commit special-cases `BuiltinImplSource::Trivial` to avoid this - that might not be precisely the correct fix but it is unclear what that would be. Without this, the following tests fail: - tests/ui/associated-type-bounds/cant-see-copy-bound-from-child-rigid.rs#next - tests/ui/associated-types/defaults-suitability.rs#next - tests/ui/associated-types/defaults-unsound-62211-1.rs#next - tests/ui/associated-types/defaults-unsound-62211-2.rs#next - tests/ui/associated-types/imply-relevant-nested-item-bounds.rs#next - tests/ui/associated-types/imply-relevant-nested-item-bounds-2.rs#next - tests/ui/associated-types/issue-63593.rs#next - tests/ui/associated-types/issue-54108.rs#next - tests/ui/async-await/async-closures/constrained-but-no-upvars-yet.rs#next - tests/ui/async-await/async-closures/is-not-fn.rs#next - tests/ui/async-await/async-closures/is-fn.rs#next - tests/ui/async-await/async-closures/once.rs#next - tests/ui/async-await/async-closures/signature-inference-from-two-part-bound.rs#next - tests/ui/async-await/async-fn/higher-ranked-async-fn.rs#next - tests/ui/async-await/async-fn/project.rs#next - tests/ui/async-await/return-type-notation/normalizing-self-auto-trait-issue-109924.rs#next - tests/ui/async-await/send-bound-async-closure.rs#next - tests/ui/borrowck/implied-bound-from-impl-header.rs#next - tests/ui/borrowck/implied-bound-from-normalized-arg.rs#next - tests/ui/borrowck/issue-103095.rs#next - tests/ui/closures/deduce-signature/deduce-from-opaque-type-after-norm.rs#next - tests/ui/closures/deduce-signature/infer-higher-ranked-signature.rs#next - tests/ui/closures/deduce-signature/infer-signature-from-impl.rs#next - tests/ui/closures/deduce-signature/obligation-with-leaking-placeholders.rs#next - tests/ui/closures/deduce-signature/supertrait-signature-inference-issue-23012.rs#next - tests/ui/closures/supertrait-hint-references-assoc-ty.rs#next - tests/ui/coherence/coherence-overlap-negate-use-feature-gate.rs - tests/ui/coherence/occurs-check/opaques.rs#next - tests/ui/coherence/occurs-check/associated-type.rs#next - tests/ui/coherence/orphan-check-alias.rs#next - tests/ui/coherence/orphan-check-projections-nested.rs#next - tests/ui/coherence/orphan-check-projections-not-covering-ambiguity.rs#next - tests/ui/coherence/orphan-check-projections-not-covering.rs#next - tests/ui/coroutine/gen_block_iterate.rs#next - tests/ui/diagnostic_namespace/do_not_recommend/as_expression.rs#next - tests/ui/diagnostic_namespace/do_not_recommend/does_not_acccept_args.rs#next - tests/ui/diagnostic_namespace/do_not_recommend/nested.rs#next - tests/ui/diagnostic_namespace/do_not_recommend/simple.rs#next - tests/ui/diagnostic_namespace/do_not_recommend/stacked.rs#next - tests/ui/diagnostic_namespace/do_not_recommend/type_mismatch.rs#next - tests/ui/diagnostic_namespace/do_not_recommend/supress_suggestions_in_help.rs#next - tests/ui/diagnostic_namespace/do_not_recommend/with_lifetime.rs#next - tests/ui/generic-associated-types/ambig-hr-projection-issue-93340.rs#next - tests/ui/generic-associated-types/assume-gat-normalization-for-nested-goals.rs#next - tests/ui/generic-associated-types/issue-74816.rs#next - tests/ui/generic-associated-types/issue-74824.rs#next - tests/ui/generic-associated-types/issue-91883.rs#next - tests/ui/generic-associated-types/rigid-hr-projection-issue-93340.rs#next - tests/ui/generic-const-items/const-trait-impl.rs - tests/ui/higher-ranked/builtin-closure-like-bounds.rs#next - tests/ui/higher-ranked/closure-bound-codegen-ice.rs#next - tests/ui/higher-ranked/leak-check/candidate-from-env-universe-err-1.rs#next - tests/ui/higher-ranked/leak-check/candidate-from-env-universe-err-2.rs#next - tests/ui/higher-ranked/leak-check/candidate-from-env-universe-err-project.rs#next - tests/ui/higher-ranked/leak-check/leak-check-in-selection-2.rs#next - tests/ui/higher-ranked/leak-check/leak-check-in-selection-3.rs#next - tests/ui/higher-ranked/leak-check/leak-check-in-selection-4-hr-nested.rs#next - tests/ui/higher-ranked/leak-check/leak-check-in-selection-5-ambig.rs#next - tests/ui/higher-ranked/leak-check/leak-check-in-selection-6-ambig-unify.rs#next - tests/ui/higher-ranked/trait-bounds/future.rs#next - tests/ui/higher-ranked/trait-bounds/rigid-equate-projections-in-higher-ranked-fn-signature.rs#next - tests/ui/higher-ranked/trait-bounds/normalize-under-binder/norm-before-method-resolution-opaque-type.rs#next - tests/ui/impl-trait/auto-trait-selection-freeze.rs#next - tests/ui/impl-trait/auto-trait-selection.rs#next - tests/ui/impl-trait/call_method_on_inherent_impl.rs#next - tests/ui/impl-trait/call_method_ambiguous.rs#next - tests/ui/impl-trait/dyn-trait-elided-two-inputs-ref-assoc.rs#next - tests/ui/impl-trait/hidden-type-is-opaque-2.rs#next - tests/ui/impl-trait/in-trait/alias-bounds-when-not-wf.rs - tests/ui/impl-trait/in-trait/default-body-with-rpit.rs#next - tests/ui/impl-trait/in-trait/dyn-compatibility-sized.rs#next - tests/ui/impl-trait/in-trait/false-positive-predicate-entailment-error.rs#next - tests/ui/impl-trait/in-trait/refine-normalize.rs#next - tests/ui/impl-trait/issue-103181-1.rs#next - tests/ui/impl-trait/method-resolution2.rs#next - tests/ui/impl-trait/recursive-bound-eval.rs#next - tests/ui/impl-trait/unsized_coercion.rs#next - tests/ui/issues/issue-15734.rs#next - tests/ui/methods/fulfillment-disqualifies-method.rs#next - tests/ui/methods/leak-check-disquality.rs#next - tests/ui/nll/check-normalized-sig-for-wf.rs#next - tests/ui/panics/abort-on-panic.rs#next - tests/ui/specialization/source-impl-requires-constraining-predicates.rs#next - tests/ui/specialization/source-impl-requires-constraining-predicates-ambig.rs#next - tests/ui/specialization/specialization-default-items-drop-coherence.rs#next - tests/ui/traits/const-traits/assoc-type-const-bound-usage-0.rs#next - tests/ui/traits/const-traits/assoc-type-const-bound-usage-fail-2.rs#next - tests/ui/traits/const-traits/assoc-type-const-bound-usage-fail.rs#next - tests/ui/traits/const-traits/assoc-type.rs#next - tests/ui/traits/const-traits/call-const-in-tilde-const.rs - tests/ui/traits/const-traits/call-generic-method-dup-bound.rs - tests/ui/traits/const-traits/call-generic-method-nonconst.rs - tests/ui/traits/const-traits/const-bound-in-host.rs - tests/ui/traits/const-traits/const-bound-on-not-const-associated-fn.rs - tests/ui/traits/const-traits/const-cond-for-rpitit.rs - tests/ui/traits/const-traits/const-drop-fail.rs#new_precise - tests/ui/traits/const-traits/const-drop-fail.rs#new_stock - tests/ui/traits/const-traits/const-drop.rs#precise - tests/ui/traits/const-traits/const-drop.rs#stock - tests/ui/traits/const-traits/const-in-closure.rs - tests/ui/traits/const-traits/const-opaque.rs#no - tests/ui/traits/const-traits/const-impl-trait.rs - tests/ui/traits/const-traits/const-opaque.rs#yes - tests/ui/traits/const-traits/dont-ice-on-const-pred-for-bounds.rs - tests/ui/traits/const-traits/infer-fallback.rs - tests/ui/traits/const-traits/issue-92230-wf-super-trait-env.rs - tests/ui/traits/const-traits/item-bound-entailment-fails.rs - tests/ui/traits/const-traits/item-bound-entailment.rs - tests/ui/traits/const-traits/minicore-drop-without-feature-gate.rs#no - tests/ui/traits/const-traits/minicore-deref-fail.rs - tests/ui/traits/const-traits/minicore-const-fn-early-bound.rs - tests/ui/traits/const-traits/minicore-drop-fail.rs - tests/ui/traits/const-traits/minicore-drop-without-feature-gate.rs#yes - tests/ui/traits/const-traits/minicore-fn-fail.rs - tests/ui/traits/const-traits/minicore-works.rs - tests/ui/traits/const-traits/predicate-entailment-passes.rs - tests/ui/traits/const-traits/predicate-entailment-fails.rs - tests/ui/traits/const-traits/super-traits-fail-2.rs#nn - tests/ui/traits/const-traits/super-traits-fail-2.rs#ny - tests/ui/traits/const-traits/staged-api.rs - tests/ui/traits/const-traits/super-traits-fail-3.rs#nnn - tests/ui/traits/const-traits/super-traits-fail-2.rs#yn - tests/ui/traits/const-traits/super-traits-fail-2.rs#yy - tests/ui/traits/const-traits/super-traits-fail-3.rs#nny - tests/ui/traits/const-traits/super-traits-fail-3.rs#nyy - tests/ui/traits/const-traits/super-traits-fail-3.rs#nyn - tests/ui/traits/const-traits/super-traits-fail-3.rs#yyy - tests/ui/traits/const-traits/super-traits-fail-3.rs#ynn - tests/ui/traits/const-traits/super-traits-fail-3.rs#yyn - tests/ui/traits/const-traits/super-traits-fail-3.rs#yny - tests/ui/traits/const-traits/super-traits.rs - tests/ui/traits/const-traits/tilde-const-assoc-fn-in-trait-impl.rs - tests/ui/traits/const-traits/tilde-const-inherent-assoc-const-fn.rs - tests/ui/traits/const-traits/tilde-const-in-struct-args.rs - tests/ui/traits/const-traits/trait-where-clause-self-referential.rs - tests/ui/traits/const-traits/trait-where-clause-const.rs - tests/ui/traits/const-traits/unsatisfied-const-trait-bound.rs - tests/ui/traits/negative-bounds/opaque-type-unsatisfied-bound.rs - tests/ui/traits/negative-impls/negative-impl-normalizes-to.rs#next - tests/ui/traits/next-solver/alias-bound-preference.rs#next - tests/ui/traits/next-solver/alias-relate/alias_eq_cant_be_furthur_normalized.rs - tests/ui/traits/next-solver/alias-eq-in-canonical-response.rs - tests/ui/traits/next-solver/alias-bound-unsound.rs - tests/ui/traits/next-solver/alias-relate/alias_eq_simple.rs - tests/ui/traits/next-solver/alias-relate/opaque-hidden-ty-is-rigid-alias.rs - tests/ui/traits/next-solver/alias-sub.rs - tests/ui/traits/next-solver/alias-relate/deeply-nested-no-hang.rs - tests/ui/traits/next-solver/assembly/ambig-projection-self-is-ambig.rs - tests/ui/traits/next-solver/assembly/assemble-normalizing-self-ty-impl-ambiguity.rs - tests/ui/traits/next-solver/assembly/candidates-equal-modulo-norm-1.rs - tests/ui/traits/next-solver/async.rs#fail - tests/ui/traits/next-solver/assembly/candidates-equal-modulo-norm-2.rs - tests/ui/traits/next-solver/async.rs#pass - tests/ui/traits/next-solver/assembly/param-env-alias-bound-conflict.rs - tests/ui/traits/next-solver/auto-with-drop_tracking_mir.rs#pass - tests/ui/traits/next-solver/assembly/runaway-impl-candidate-selection.rs - tests/ui/traits/next-solver/auto-with-drop_tracking_mir.rs#fail - tests/ui/traits/next-solver/canonical/effect-var.rs - tests/ui/traits/next-solver/canonical/int-var-eq-in-response.rs - tests/ui/traits/next-solver/builtin-fn-must-return-sized.rs - tests/ui/traits/next-solver/closure-inference-guidance.rs - tests/ui/traits/next-solver/closure-signature-inference-2.rs - tests/ui/traits/next-solver/closure-signature-inference-hr-ambig-alias-naming-self.rs#next - tests/ui/traits/next-solver/coerce-ambig-alias-to-rigid-alias.rs - tests/ui/traits/next-solver/coherence/trait_ref_is_knowable-norm-overflow.rs - tests/ui/traits/next-solver/coherence/trait_ref_is_knowable-normalization-1.rs - tests/ui/traits/next-solver/coherence/trait_ref_is_knowable-normalization-3.rs - tests/ui/traits/next-solver/coherence/trait_ref_is_knowable-normalization-2.rs - tests/ui/traits/next-solver/const-param-placeholder.rs#pass - tests/ui/traits/next-solver/const-param-placeholder.rs#fail - tests/ui/traits/next-solver/constrain-alias-goals-in-unsize.rs - tests/ui/traits/next-solver/coroutine.rs#pass - tests/ui/traits/next-solver/coroutine.rs#fail - tests/ui/traits/next-solver/cycles/coinduction/fixpoint-exponential-growth.rs - tests/ui/traits/next-solver/cycles/double-cycle-inductive-coinductive.rs - tests/ui/traits/next-solver/cycles/inductive-cycle-but-err.rs - tests/ui/traits/next-solver/cycles/inductive-cycle-but-ok.rs - tests/ui/traits/next-solver/cycles/inductive-fixpoint-hang.rs - tests/ui/traits/next-solver/cycles/leak-check-coinductive-cycle.rs - tests/ui/traits/next-solver/cycles/inductive-not-on-stack.rs - tests/ui/traits/next-solver/cycles/mixed-cycles-1.rs - tests/ui/traits/next-solver/cycles/mixed-cycles-2.rs - tests/ui/traits/next-solver/cycles/provisional-cache-impacts-behavior.rs - tests/ui/traits/next-solver/cycles/provisional-result-done.rs - tests/ui/traits/next-solver/cycles/cycle-modulo-ambig-aliases.rs - tests/ui/traits/next-solver/destruct.rs - tests/ui/traits/next-solver/diagnostics/alias_relate_error_uses_structurally_normalize.rs - tests/ui/traits/next-solver/diagnostics/deeply-normalize-type-expectation.rs - tests/ui/traits/next-solver/diagnostics/coerce-in-may-coerce.rs - tests/ui/traits/next-solver/diagnostics/point-at-failing-nested.rs - tests/ui/traits/next-solver/diagnostics/projection-trait-ref.rs - tests/ui/traits/next-solver/diagnostics/dont-pick-fnptr-bound-as-leaf.rs#next - tests/ui/traits/next-solver/dont-loop-fulfill-on-region-constraints.rs - tests/ui/traits/next-solver/elaborate-item-bounds.rs - tests/ui/traits/next-solver/dont-normalize-proj-with-error.rs - tests/ui/traits/next-solver/dyn-incompatibility.rs - tests/ui/traits/next-solver/fn-trait-closure.rs - tests/ui/traits/next-solver/fn-trait.rs - tests/ui/traits/next-solver/generalize/bivariant-alias.rs#next - tests/ui/traits/next-solver/generalize/equating-projection-cyclically.rs - tests/ui/traits/next-solver/generalize/hr-alias-non-hr-alias-self-ty-1.rs#next - tests/ui/traits/next-solver/generalize/hr-alias-universe-lowering-ambiguity.rs - tests/ui/traits/next-solver/generalize/occurs-check-nested-alias.rs#next - tests/ui/traits/next-solver/generalize/instantiate-canonical-occurs-check-failure.rs - tests/ui/traits/next-solver/generalize/hr-alias-non-hr-alias-self-ty-2.rs#next - tests/ui/traits/next-solver/higher-ranked-dyn-bounds.rs - tests/ui/traits/next-solver/issue-118950-root-region.rs - tests/ui/traits/next-solver/iter-filter-projection.rs - tests/ui/traits/next-solver/lazy-nested-obligations-2.rs - tests/ui/traits/next-solver/lazy-nested-obligations-3.rs - tests/ui/traits/next-solver/nested-obligations-with-bound-vars-gat.rs - tests/ui/traits/next-solver/nested-alias-bound.rs - tests/ui/traits/next-solver/method/path_lookup_wf_constraints.rs - tests/ui/traits/next-solver/non-wf-ret.rs - tests/ui/traits/next-solver/normalization-shadowing/alias-bound-shadowed-by-env.rs - tests/ui/traits/next-solver/normalization-shadowing/discard-impls-shadowed-by-env-1.rs - tests/ui/traits/next-solver/normalization-shadowing/discard-impls-shadowed-by-env-3.rs - tests/ui/traits/next-solver/normalization-shadowing/discard-impls-shadowed-by-env-2.rs#next - tests/ui/traits/next-solver/normalization-shadowing/param-env-impl-conflict.rs - tests/ui/traits/next-solver/normalization-shadowing/normalizes_to_ignores_unnormalizable_candidate.rs - tests/ui/traits/next-solver/normalization-shadowing/param-candidate-shadows-project.rs - tests/ui/traits/next-solver/normalize-in-implied_outlives_bounds.rs - tests/ui/traits/next-solver/normalize/normalize-param-env-3.rs - tests/ui/traits/next-solver/normalize/normalize-param-env-1.rs - tests/ui/traits/next-solver/normalize/normalize-async-closure-in-trait.rs - tests/ui/traits/next-solver/normalize/normalize-param-env-2.rs - tests/ui/traits/next-solver/normalize/normalize-param-env-4.rs#next - tests/ui/traits/next-solver/normalize/normalize-rcvr-for-inherent.rs - tests/ui/traits/next-solver/normalize/normalize-path-for-method.rs - tests/ui/traits/next-solver/normalize/normalize-region-obligations.rs#hrtb - tests/ui/traits/next-solver/normalize/normalize-region-obligations.rs#normalize_obligation - tests/ui/traits/next-solver/normalize/normalize-region-obligations.rs#normalize_param_env - tests/ui/traits/next-solver/normalize/normalize-type-outlives-in-param-env.rs - tests/ui/traits/next-solver/normalize/ambig-goal-infer-in-type-oulives.rs - tests/ui/traits/next-solver/normalize/param-env-trait-candidate-2.rs - tests/ui/traits/next-solver/normalize/two-projection-param-candidates-are-ambiguous.rs - tests/ui/traits/next-solver/opaques/dont-remap-tait-substs.rs - tests/ui/traits/next-solver/opaques/select-alias-bound-as-param.rs - tests/ui/traits/next-solver/opportunistic-region-resolve.rs - tests/ui/traits/next-solver/opaques/no-define-in-wf-check.rs#next - tests/ui/traits/next-solver/overflow/exponential-trait-goals.rs - tests/ui/traits/next-solver/normalize/param-env-trait-candidate-1.rs - tests/ui/traits/next-solver/overflow/global-cache.rs - tests/ui/traits/next-solver/overflow/nalgebra-hang.rs#next - tests/ui/traits/next-solver/overflow/recursive-self-normalization.rs - tests/ui/traits/next-solver/overflow/recursive-self-normalization-2.rs - tests/ui/traits/next-solver/pointee.rs - tests/ui/traits/next-solver/pointer-like.rs - tests/ui/traits/next-solver/prefer-candidate-no-constraints.rs - tests/ui/traits/next-solver/prefer-param-env-on-ambiguity.rs - tests/ui/traits/next-solver/projection-discr-kind.rs - tests/ui/traits/next-solver/specialization-unconstrained.rs - tests/ui/traits/next-solver/specialization-transmute.rs - tests/ui/traits/next-solver/typeck/guide-ctors.rs - tests/ui/traits/next-solver/try-example.rs - tests/ui/traits/next-solver/typeck/normalize-in-upvar-collection.rs - tests/ui/traits/next-solver/typeck/resolve-before-checking-builtin-ptr.rs - tests/ui/traits/next-solver/typeck/resolve-before-checking-never.rs - tests/ui/traits/next-solver/typeck/structurally-resolve-in-probe_adt.rs - tests/ui/traits/next-solver/typeck/resolve-expectations.rs - tests/ui/traits/next-solver/winnow-specializing-impls.rs - tests/ui/traits/non_lifetime_binders/unifying-placeholders-in-query-response.rs#next - tests/ui/traits/normalize-supertrait.rs#next - tests/ui/traits/pointee-normalize-equate.rs#next - tests/ui/traits/reservation-impl/non-lattice-ok.rs#next - tests/ui/traits/trait-upcasting/impossible-method-modulo-binders-2.rs#next - tests/ui/traits/trait-upcasting/impossible-method-modulo-binders.rs#next - tests/ui/traits/trait-upcasting/normalization.rs#next - tests/ui/traits/trait-upcasting/prefer-lower-candidates.rs#next - tests/ui/traits/winnowing/global-non-global-env-1.rs#next - tests/ui/traits/winnowing/global-non-global-env-3.rs#next - tests/ui/traits/winnowing/global-non-global-env-2.rs#next - tests/ui/traits/winnowing/global-non-global-env-4.rs#next - tests/ui/transmutability/primitives/bool.rs#next - tests/ui/transmutability/primitives/unit.rs#next - tests/ui/type-alias-impl-trait/assoc-type-const.rs#next - tests/ui/type-alias-impl-trait/issue-76202-trait-impl-for-tait.rs#next - tests/ui/type-alias-impl-trait/issue-78450.rs#next - tests/ui/type-alias-impl-trait/issue-84660-unsoundness.rs#next - tests/ui/type-alias-impl-trait/method_resolution3.rs#next - tests/ui/type-alias-impl-trait/method_resolution4.rs#next - tests/ui/type-alias-impl-trait/nested_inference_failure.rs#next - tests/ui/type-alias-impl-trait/normalize-hidden-types.rs#next - tests/ui/type-alias-impl-trait/rpit_tait_equality_in_canonical_query.rs#next - tests/ui/typeck/bad-index-due-to-nested.rs#next - tests/ui/typeck/issue-116864.rs - tests/ui/unsized/issue-75899.rs#next - tests/ui/transmutability/primitives/numbers.rs#next - tests/ui/wf/wf-in-where-clause-static.rs#next

When evaluating a host effect predicate on an opaque in `addr2line`, running `evaluate_host_effect_from_bounds` first would create an opaque type which wouldn't be removed from opaque type storage and that would cause a delayed bug - creating a small reproduction of this has been tricky but if `builtin_impls` checking is moved below `evaluate_host_effect_from_item_bounds` then it will trigger. The correct fix is probably to work out why the opaque type is created and then not used, but this was particularly tricky. Given that this root cause wasn't identified, writing a test case proved difficult.

Now that sizedness traits have been made const, pretty printing of impl headers should print the constness of the sizedness trait.

These traits are now const and that needs to be reflected in their printing in opaques.

This causes a bunch of tests to fail and everything appears to work without it: - tests/ui/coherence/orphan-check-opaque-types-not-covering.rs - tests/ui/coherence/orphan-check-projections-covering.rs - tests/ui/coherence/orphan-check-projections-unsat-bounds.rs#classic - tests/ui/coherence/orphan-check-weak-aliases-not-covering.rs - tests/ui/coherence/orphan-check-weak-aliases-covering.rs#classic - tests/ui/specialization/issue-43037.rs#negative - tests/ui/specialization/issue-43037.rs#current - tests/ui/type-alias-impl-trait/coherence.rs#classic

These tests just need blessing, they don't have any interesting behaviour changes.

Its strange to have to import sizedness traits, so add them to the prelude in the next edition.

Implement the edition migration required by these changes - `?Sized` becoming `const MetaSized`, `Sized` becoming `const Sized` and a `const MetaSized` supertrait.

This should have been included in rust-lang#137606.

This test no longer crashes the compiler as `Box` no longer accepts `PointeeSized`-types. It eventually could, but not because of `Deref::Target` currently, so this doesn't fail anymore and there wasn't an obvious to add new types to make it continue to fail because `Deref` is special.

Some rustdoc tests are `no_core` and need to have `MetaSized` and `PointeeSized` added to them. One or two tests used extern types and needed further relaxations to work as before.

These should never be shown to users at the moment.

One clippy test is `no_core` and needs to have `MetaSized` and `PointeeSized` added to it.

Existing lints that had special-casing for `Sized` predicates ought to have these same special cases applied to `MetaSized` predicates.

Unexpected Clippy lint triggering is fixed in previous commits but is necessary for `cfg(bootstrap)`.

It's unclear why this change in miri is necessary.

These error messages include lines of the standard library which have changed and so need updated.

Unstability is propagated when staged API is not used and `-Zforce-unstable-if-unmarked` is, but this only happened for const functions not const traits, which ends up being an issue for some of the minicore tests of codegen backends when introducing the sizedness traits.

As in many previous commits, adding the new traits and constness to the minicore.

bors · 2025-03-07T17:51:45Z

☔ The latest upstream changes (presumably #138155) made this pull request unmergeable. Please resolve the merge conflicts.

fee1-dead

had an initial look until the "library/compiler" commit. Will look more into it later.

fee1-dead · 2025-03-08T08:43:02Z

compiler/rustc_middle/src/ty/adt.rs

@@ -625,6 +633,24 @@ impl<'tcx> AdtDef<'tcx> {
    pub fn sized_constraint(self, tcx: TyCtxt<'tcx>) -> Option<ty::EarlyBinder<'tcx, Ty<'tcx>>> {
        if self.is_struct() { tcx.adt_sized_constraint(self.did()) } else { None }
    }
+
+    /// Returns a type such that `Self: MetaSized` if and only if that type is `MetaSized`,
+    /// or `None` if the type is always metasized.


Should capitalize MetaSized here

fee1-dead · 2025-03-08T08:43:09Z

compiler/rustc_middle/src/ty/adt.rs

+    }
+
+    /// Returns a type such that `Self: PointeeSized` if and only if that type is `PointeeSized`,
+    /// or `None` if the type is always pointeesized.


same as above

fee1-dead · 2025-03-08T08:43:48Z

compiler/rustc_middle/src/ty/adt.rs

+
+    /// Returns a type such that `Self: PointeeSized` if and only if that type is `PointeeSized`,
+    /// or `None` if the type is always pointeesized.
+    pub fn pointeesized_constraint(


Hmm. I think this should be named as pointee_sized_constraint for better readability? Same goes for meta_sized_constraint

fee1-dead · 2025-03-08T08:46:18Z

compiler/rustc_next_trait_solver/src/solve/assembly/structural_traits.rs

+        // impl {Meta,Pointee,}Sized for str, [T], dyn Trait
+        ty::Str | ty::Slice(_) | ty::Dynamic(..) => match sizedness {
+            Sizedness::Sized => Err(NoSolution),
+            Sizedness::MetaSized | Sizedness::PointeeSized => Ok(ty::Binder::dummy(vec![])),
+        },


Is something like [str] PointeeSized? Might be a dumb question since that's probably not well-formed?

It is, everything is PointeeSized.

fee1-dead · 2025-03-08T08:47:26Z

compiler/rustc_next_trait_solver/src/solve/assembly/structural_traits.rs

+        // impl {Meta,}Sized for ()
+        // impl {Meta,}Sized for (T1, T2, .., Tn) where Tn: {Meta,}Sized if n >= 1


are these missing the `{Meta,Pointee,} or intentional?

fee1-dead · 2025-03-08T09:31:06Z

compiler/rustc_type_ir/src/solve/mod.rs

+
+/// Which sizedness trait - `Sized`, `MetaSized` or `PointeeSized`?
+#[derive(Copy, Clone, Debug)]
+pub enum Sizedness {


I'd prefer it to be named as SizedTraitKind.

fee1-dead · 2025-03-08T09:33:32Z

tests/ui/sized-hierarchy/impls.rs

+    needs_pointeesized::<([u8], [u8])>();
+    //~^ ERROR the size for values of type `[u8]` cannot be known at compilation time
+
+    // tuple w/ all elements unsized


should be PointeeSized

fee1-dead · 2025-03-08T09:33:40Z

tests/ui/sized-hierarchy/impls.rs

+    needs_metasized::<(u32, [u8])>();
+    needs_pointeesized::<(u32, [u8])>();
+
+    // tuple w/ last element unsized


PointeeSized

fee1-dead · 2025-03-08T09:37:20Z

tests/ui/sized-hierarchy/impls.rs

+    struct StructAllFieldsMetaSized { x: [u8], y: [u8] }
+    //~^ ERROR the size for values of type `[u8]` cannot be known at compilation time
+    needs_sized::<StructAllFieldsMetaSized>();
+    //~^ ERROR the size for values of type `[u8]` cannot be known at compilation time
+    needs_metasized::<StructAllFieldsMetaSized>();


I'm a little confused as to what this achieves with this kind of type implementing MetaSized, although that is what is described in the RfC..

fee1-dead · 2025-03-08T09:59:12Z

library/core/src/cmp.rs

+/// Trait for comparisons using the equality operator.
+///
+/// Implementing this trait for types provides the `==` and `!=` operators for
+/// those types.


You don't need docs for cfg(bootstrap)ed items.

This was referenced Mar 3, 2025

Hierarchy of Sized traits rust-lang/rfcs#3729

Open

SVE and SME on AArch64 rust-lang/rust-project-goals#270

Open

This comment has been minimized.

Sign in to view

rustbot added the T-bootstrap Relevant to the bootstrap subteam: Rust's build system (x.py and src/bootstrap) label Mar 3, 2025

This comment has been minimized.

Sign in to view

davidtwco force-pushed the sized-hierarchy branch from 73f1b4f to 7f509ab Compare March 3, 2025 16:29

This comment has been minimized.

Sign in to view

davidtwco force-pushed the sized-hierarchy branch from 2537bfb to b31fd85 Compare March 3, 2025 18:04

workingjubilee reviewed Mar 3, 2025

View reviewed changes

compiler/rustc_type_ir/src/flags.rs Outdated Show resolved Hide resolved

compiler/rustc_passes/src/check_attr.rs Outdated Show resolved Hide resolved

tests/ui/target-feature/tied-features-no-implication-1.rs Show resolved Hide resolved

traviscross added the T-lang Relevant to the language team, which will review and decide on the PR/issue. label Mar 4, 2025

fee1-dead self-assigned this Mar 4, 2025

This comment was marked as resolved.

Sign in to view

davidtwco force-pushed the sized-hierarchy branch from b31fd85 to 2beed43 Compare March 6, 2025 10:21

davidtwco force-pushed the sized-hierarchy branch from 2beed43 to 839b844 Compare March 6, 2025 14:22

upstream: rust-lang#137606

a5ca7ee

davidtwco added 26 commits March 6, 2025 17:47

feature/passes: add rustc_non_const_sized

c75d064

While there are no non-`const Sized` types in the language, this testing attribute will make it possible to force this for testing purposes.

library/trait_sel: make {Meta,}Sized const

cccc9a1

Makes `{Meta,}Sized` into const traits and implements them in the trait solvers so everything is `Sized` except ADTs with the `rustc_non_const_sized` attribute.

aux: add const {Meta,}Sized to minicore

fc73927

Now that `MetaSized` and `Sized` are const traits, this must be reflected in minicore.

tests: const {Meta,}Sized in non-minicore

497f211

Now that `MetaSized` and `Sized` are const traits, this must be reflected in tests without minicore which define these language items.

hir_analysis: change {Meta,}Sized bounds to const

6c60d5c

With const sizedness traits, the correct migration is from `Sized` to `const Sized` and from `?Sized` to `const MetaSized`.

hir_analysis: specialisation w/ host effect preds

a269792

There are some uses of `min_specialization` in the standard library which now see `HostEffectPredicate` - this shouldn't prevent specialization for the sizedness marker traits that are now const.

trait_sel: const {Meta,}Sized in impl headers

816c625

Now that sizedness traits have been made const, pretty printing of impl headers should print the constness of the sizedness trait.

middle: const {Meta,Pointee}Sized in opaques

a487cfc

These traits are now const and that needs to be reflected in their printing in opaques.

tests: bless remaining tests

69cfa2a

These tests just need blessing, they don't have any interesting behaviour changes.

library: add {Pointee,Meta}Sized to next prelude

6455c8a

Its strange to have to import sizedness traits, so add them to the prelude in the next edition.

hir_analysis: edition migration

125d861

Implement the edition migration required by these changes - `?Sized` becoming `const MetaSized`, `Sized` becoming `const Sized` and a `const MetaSized` supertrait.

lint-docs: add future edition group

c013e82

This should have been included in rust-lang#137606.

rustdoc: update no_core tests w/ new traits

f9e8ad9

Some rustdoc tests are `no_core` and need to have `MetaSized` and `PointeeSized` added to them. One or two tests used extern types and needed further relaxations to work as before.

rustdoc: skip MetaSized bounds

4e76b7e

These should never be shown to users at the moment.

clippy: update no_core tests w/ new traits

dc5f1ae

One clippy test is `no_core` and needs to have `MetaSized` and `PointeeSized` added to it.

clippy: add MetaSized conditions

2089f2e

Existing lints that had special-casing for `Sized` predicates ought to have these same special cases applied to `MetaSized` predicates.

bootstrap: address lint failures

047ea07

Unexpected Clippy lint triggering is fixed in previous commits but is necessary for `cfg(bootstrap)`.

miri: fix build

6731723

It's unclear why this change in miri is necessary.

miri: bless tests

5e4ef35

These error messages include lines of the standard library which have changed and so need updated.

cranelift/gcc: add sizedness traits to minicore

1874a3d

As in many previous commits, adding the new traits and constness to the minicore.

This comment was marked as resolved.

Sign in to view

davidtwco force-pushed the sized-hierarchy branch from 839b844 to 1874a3d Compare March 7, 2025 14:03

fee1-dead reviewed Mar 8, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Sized Hierarchy #137944

Sized Hierarchy #137944

davidtwco commented Mar 3, 2025 •

edited

Loading

This comment has been minimized.

davidtwco commented Mar 3, 2025 •

edited

Loading

This comment has been minimized.

This comment has been minimized.

traviscross commented Mar 4, 2025

tmiasko commented Mar 4, 2025

This comment was marked as resolved.

davidtwco commented Mar 6, 2025

This comment was marked as resolved.

bors commented Mar 7, 2025

fee1-dead left a comment

fee1-dead Mar 8, 2025

fee1-dead Mar 8, 2025

fee1-dead Mar 8, 2025

fee1-dead Mar 8, 2025

davidtwco Mar 10, 2025

fee1-dead Mar 8, 2025

fee1-dead Mar 8, 2025

fee1-dead Mar 8, 2025

fee1-dead Mar 8, 2025

fee1-dead Mar 8, 2025

fee1-dead Mar 8, 2025

		// impl {Meta,}Sized for ()
		// impl {Meta,}Sized for (T1, T2, .., Tn) where Tn: {Meta,}Sized if n >= 1

Sized Hierarchy #137944

Are you sure you want to change the base?

Sized Hierarchy #137944

Conversation

davidtwco commented Mar 3, 2025 • edited Loading

This comment has been minimized.

davidtwco commented Mar 3, 2025 • edited Loading

This comment has been minimized.

This comment has been minimized.

traviscross commented Mar 4, 2025

tmiasko commented Mar 4, 2025

This comment was marked as resolved.

davidtwco commented Mar 6, 2025

This comment was marked as resolved.

bors commented Mar 7, 2025

fee1-dead 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

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

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

davidtwco commented Mar 3, 2025 •

edited

Loading

davidtwco commented Mar 3, 2025 •

edited

Loading