//! The new trait solver, currently still WIP. //! //! As a user of the trait system, you can use `TyCtxt::evaluate_goal` to //! interact with this solver. //! //! For a high-level overview of how this solver works, check out the relevant //! section of the rustc-dev-guide. //! //! FIXME(@lcnr): Write that section. If you read this before then ask me //! about it on zulip. // FIXME: Instead of using `infcx.canonicalize_query` we have to add a new routine which // preserves universes and creates a unique var (in the highest universe) for each // appearance of a region. // FIXME: uses of `infcx.at` need to enable deferred projection equality once that's implemented. use std::mem; use rustc_hir::def_id::DefId; use rustc_infer::infer::canonical::{Canonical, CanonicalVarValues}; use rustc_infer::infer::{InferCtxt, InferOk, TyCtxtInferExt}; use rustc_infer::traits::query::NoSolution; use rustc_infer::traits::Obligation; use rustc_middle::traits::solve::{ExternalConstraints, ExternalConstraintsData}; use rustc_middle::ty::{self, Ty, TyCtxt}; use rustc_middle::ty::{ CoercePredicate, RegionOutlivesPredicate, SubtypePredicate, ToPredicate, TypeOutlivesPredicate, }; use rustc_span::DUMMY_SP; use crate::solve::search_graph::OverflowHandler; use crate::traits::ObligationCause; mod assembly; mod canonical; mod eval_ctxt; mod fulfill; mod project_goals; mod search_graph; mod trait_goals; pub use eval_ctxt::EvalCtxt; pub use fulfill::FulfillmentCtxt; /// A goal is a statement, i.e. `predicate`, we want to prove /// given some assumptions, i.e. `param_env`. /// /// Most of the time the `param_env` contains the `where`-bounds of the function /// we're currently typechecking while the `predicate` is some trait bound. #[derive(Debug, PartialEq, Eq, Clone, Copy, Hash, TypeFoldable, TypeVisitable)] pub struct Goal<'tcx, P> { param_env: ty::ParamEnv<'tcx>, predicate: P, } impl<'tcx, P> Goal<'tcx, P> { pub fn new( tcx: TyCtxt<'tcx>, param_env: ty::ParamEnv<'tcx>, predicate: impl ToPredicate<'tcx, P>, ) -> Goal<'tcx, P> { Goal { param_env, predicate: predicate.to_predicate(tcx) } } /// Updates the goal to one with a different `predicate` but the same `param_env`. fn with(self, tcx: TyCtxt<'tcx>, predicate: impl ToPredicate<'tcx, Q>) -> Goal<'tcx, Q> { Goal { param_env: self.param_env, predicate: predicate.to_predicate(tcx) } } } impl<'tcx, P> From> for Goal<'tcx, P> { fn from(obligation: Obligation<'tcx, P>) -> Goal<'tcx, P> { Goal { param_env: obligation.param_env, predicate: obligation.predicate } } } #[derive(Debug, PartialEq, Eq, Clone, Copy, Hash, TypeFoldable, TypeVisitable)] pub struct Response<'tcx> { pub var_values: CanonicalVarValues<'tcx>, /// Additional constraints returned by this query. pub external_constraints: ExternalConstraints<'tcx>, pub certainty: Certainty, } trait CanonicalResponseExt { fn has_no_inference_or_external_constraints(&self) -> bool; } impl<'tcx> CanonicalResponseExt for Canonical<'tcx, Response<'tcx>> { fn has_no_inference_or_external_constraints(&self) -> bool { self.value.external_constraints.region_constraints.is_empty() && self.value.var_values.is_identity() && self.value.external_constraints.opaque_types.is_empty() } } #[derive(Debug, PartialEq, Eq, Clone, Copy, Hash, TypeFoldable, TypeVisitable)] pub enum Certainty { Yes, Maybe(MaybeCause), } impl Certainty { pub const AMBIGUOUS: Certainty = Certainty::Maybe(MaybeCause::Ambiguity); /// When proving multiple goals using **AND**, e.g. nested obligations for an impl, /// use this function to unify the certainty of these goals pub fn unify_and(self, other: Certainty) -> Certainty { match (self, other) { (Certainty::Yes, Certainty::Yes) => Certainty::Yes, (Certainty::Yes, Certainty::Maybe(_)) => other, (Certainty::Maybe(_), Certainty::Yes) => self, (Certainty::Maybe(MaybeCause::Overflow), Certainty::Maybe(MaybeCause::Overflow)) => { Certainty::Maybe(MaybeCause::Overflow) } // If at least one of the goals is ambiguous, hide the overflow as the ambiguous goal // may still result in failure. (Certainty::Maybe(MaybeCause::Ambiguity), Certainty::Maybe(_)) | (Certainty::Maybe(_), Certainty::Maybe(MaybeCause::Ambiguity)) => { Certainty::Maybe(MaybeCause::Ambiguity) } } } } /// Why we failed to evaluate a goal. #[derive(Debug, PartialEq, Eq, Clone, Copy, Hash, TypeFoldable, TypeVisitable)] pub enum MaybeCause { /// We failed due to ambiguity. This ambiguity can either /// be a true ambiguity, i.e. there are multiple different answers, /// or we hit a case where we just don't bother, e.g. `?x: Trait` goals. Ambiguity, /// We gave up due to an overflow, most often by hitting the recursion limit. Overflow, } type CanonicalGoal<'tcx, T = ty::Predicate<'tcx>> = Canonical<'tcx, Goal<'tcx, T>>; type CanonicalResponse<'tcx> = Canonical<'tcx, Response<'tcx>>; /// The result of evaluating a canonical query. /// /// FIXME: We use a different type than the existing canonical queries. This is because /// we need to add a `Certainty` for `overflow` and may want to restructure this code without /// having to worry about changes to currently used code. Once we've made progress on this /// solver, merge the two responses again. pub type QueryResult<'tcx> = Result, NoSolution>; pub trait InferCtxtEvalExt<'tcx> { /// Evaluates a goal from **outside** of the trait solver. /// /// Using this while inside of the solver is wrong as it uses a new /// search graph which would break cycle detection. fn evaluate_root_goal( &self, goal: Goal<'tcx, ty::Predicate<'tcx>>, ) -> Result<(bool, Certainty), NoSolution>; } impl<'tcx> InferCtxtEvalExt<'tcx> for InferCtxt<'tcx> { fn evaluate_root_goal( &self, goal: Goal<'tcx, ty::Predicate<'tcx>>, ) -> Result<(bool, Certainty), NoSolution> { let mut search_graph = search_graph::SearchGraph::new(self.tcx); let result = EvalCtxt { search_graph: &mut search_graph, infcx: self, // Only relevant when canonicalizing the response. max_input_universe: ty::UniverseIndex::ROOT, var_values: CanonicalVarValues::dummy(), in_projection_eq_hack: false, } .evaluate_goal(goal); assert!(search_graph.is_empty()); result } } impl<'a, 'tcx> EvalCtxt<'a, 'tcx> { /// The entry point of the solver. /// /// This function deals with (coinductive) cycles, overflow, and caching /// and then calls [`EvalCtxt::compute_goal`] which contains the actual /// logic of the solver. /// /// Instead of calling this function directly, use either [EvalCtxt::evaluate_goal] /// if you're inside of the solver or [InferCtxtEvalExt::evaluate_root_goal] if you're /// outside of it. #[instrument(level = "debug", skip(tcx, search_graph), ret)] fn evaluate_canonical_goal( tcx: TyCtxt<'tcx>, search_graph: &'a mut search_graph::SearchGraph<'tcx>, canonical_goal: CanonicalGoal<'tcx>, ) -> QueryResult<'tcx> { // Deal with overflow, caching, and coinduction. // // The actual solver logic happens in `ecx.compute_goal`. search_graph.with_new_goal(tcx, canonical_goal, |search_graph| { let (ref infcx, goal, var_values) = tcx.infer_ctxt().build_with_canonical(DUMMY_SP, &canonical_goal); let mut ecx = EvalCtxt { infcx, var_values, max_input_universe: canonical_goal.max_universe, search_graph, in_projection_eq_hack: false, }; ecx.compute_goal(goal) }) } /// Recursively evaluates `goal`, returning whether any inference vars have /// been constrained and the certainty of the result. fn evaluate_goal( &mut self, goal: Goal<'tcx, ty::Predicate<'tcx>>, ) -> Result<(bool, Certainty), NoSolution> { let (orig_values, canonical_goal) = self.canonicalize_goal(goal); let canonical_response = EvalCtxt::evaluate_canonical_goal(self.tcx(), self.search_graph, canonical_goal)?; let has_changed = !canonical_response.value.var_values.is_identity(); let certainty = self.instantiate_and_apply_query_response( goal.param_env, orig_values, canonical_response, )?; // Check that rerunning this query with its inference constraints applied // doesn't result in new inference constraints and has the same result. // // If we have projection goals like `::Assoc == u32` we recursively // call `exists ::Assoc == U` to enable better caching. This goal // could constrain `U` to `u32` which would cause this check to result in a // solver cycle. if cfg!(debug_assertions) && has_changed && !self.in_projection_eq_hack && !self.search_graph.in_cycle() { let (_orig_values, canonical_goal) = self.canonicalize_goal(goal); let canonical_response = EvalCtxt::evaluate_canonical_goal(self.tcx(), self.search_graph, canonical_goal)?; if !canonical_response.value.var_values.is_identity() { bug!("unstable result: {goal:?} {canonical_goal:?} {canonical_response:?}"); } assert_eq!(certainty, canonical_response.value.certainty); } Ok((has_changed, certainty)) } fn compute_goal(&mut self, goal: Goal<'tcx, ty::Predicate<'tcx>>) -> QueryResult<'tcx> { let Goal { param_env, predicate } = goal; let kind = predicate.kind(); if let Some(kind) = kind.no_bound_vars() { match kind { ty::PredicateKind::Clause(ty::Clause::Trait(predicate)) => { self.compute_trait_goal(Goal { param_env, predicate }) } ty::PredicateKind::Clause(ty::Clause::Projection(predicate)) => { self.compute_projection_goal(Goal { param_env, predicate }) } ty::PredicateKind::Clause(ty::Clause::TypeOutlives(predicate)) => { self.compute_type_outlives_goal(Goal { param_env, predicate }) } ty::PredicateKind::Clause(ty::Clause::RegionOutlives(predicate)) => { self.compute_region_outlives_goal(Goal { param_env, predicate }) } ty::PredicateKind::Clause(ty::Clause::ConstArgHasType(ct, ty)) => { self.compute_const_arg_has_type_goal(Goal { param_env, predicate: (ct, ty) }) } ty::PredicateKind::Subtype(predicate) => { self.compute_subtype_goal(Goal { param_env, predicate }) } ty::PredicateKind::Coerce(predicate) => { self.compute_coerce_goal(Goal { param_env, predicate }) } ty::PredicateKind::ClosureKind(def_id, substs, kind) => self .compute_closure_kind_goal(Goal { param_env, predicate: (def_id, substs, kind), }), ty::PredicateKind::ObjectSafe(trait_def_id) => { self.compute_object_safe_goal(trait_def_id) } ty::PredicateKind::WellFormed(arg) => { self.compute_well_formed_goal(Goal { param_env, predicate: arg }) } ty::PredicateKind::Ambiguous => self.make_canonical_response(Certainty::AMBIGUOUS), // FIXME: implement these predicates :) ty::PredicateKind::ConstEvaluatable(_) | ty::PredicateKind::ConstEquate(_, _) => { self.make_canonical_response(Certainty::Yes) } ty::PredicateKind::TypeWellFormedFromEnv(..) => { bug!("TypeWellFormedFromEnv is only used for Chalk") } ty::PredicateKind::AliasEq(lhs, rhs) => { self.compute_alias_eq_goal(Goal { param_env, predicate: (lhs, rhs) }) } } } else { let kind = self.infcx.instantiate_binder_with_placeholders(kind); let goal = goal.with(self.tcx(), ty::Binder::dummy(kind)); let (_, certainty) = self.evaluate_goal(goal)?; self.make_canonical_response(certainty) } } fn compute_type_outlives_goal( &mut self, goal: Goal<'tcx, TypeOutlivesPredicate<'tcx>>, ) -> QueryResult<'tcx> { let ty::OutlivesPredicate(ty, lt) = goal.predicate; self.infcx.register_region_obligation_with_cause(ty, lt, &ObligationCause::dummy()); self.make_canonical_response(Certainty::Yes) } fn compute_region_outlives_goal( &mut self, goal: Goal<'tcx, RegionOutlivesPredicate<'tcx>>, ) -> QueryResult<'tcx> { self.infcx.region_outlives_predicate( &ObligationCause::dummy(), ty::Binder::dummy(goal.predicate), ); self.make_canonical_response(Certainty::Yes) } fn compute_coerce_goal( &mut self, goal: Goal<'tcx, CoercePredicate<'tcx>>, ) -> QueryResult<'tcx> { self.compute_subtype_goal(Goal { param_env: goal.param_env, predicate: SubtypePredicate { a_is_expected: false, a: goal.predicate.a, b: goal.predicate.b, }, }) } fn compute_subtype_goal( &mut self, goal: Goal<'tcx, SubtypePredicate<'tcx>>, ) -> QueryResult<'tcx> { if goal.predicate.a.is_ty_var() && goal.predicate.b.is_ty_var() { // FIXME: Do we want to register a subtype relation between these vars? // That won't actually reflect in the query response, so it seems moot. self.make_canonical_response(Certainty::AMBIGUOUS) } else { let InferOk { value: (), obligations } = self .infcx .at(&ObligationCause::dummy(), goal.param_env) .sub(goal.predicate.a, goal.predicate.b)?; self.evaluate_all_and_make_canonical_response( obligations.into_iter().map(|pred| pred.into()).collect(), ) } } fn compute_closure_kind_goal( &mut self, goal: Goal<'tcx, (DefId, ty::SubstsRef<'tcx>, ty::ClosureKind)>, ) -> QueryResult<'tcx> { let (_, substs, expected_kind) = goal.predicate; let found_kind = substs.as_closure().kind_ty().to_opt_closure_kind(); let Some(found_kind) = found_kind else { return self.make_canonical_response(Certainty::AMBIGUOUS); }; if found_kind.extends(expected_kind) { self.make_canonical_response(Certainty::Yes) } else { Err(NoSolution) } } fn compute_object_safe_goal(&mut self, trait_def_id: DefId) -> QueryResult<'tcx> { if self.tcx().check_is_object_safe(trait_def_id) { self.make_canonical_response(Certainty::Yes) } else { Err(NoSolution) } } fn compute_well_formed_goal( &mut self, goal: Goal<'tcx, ty::GenericArg<'tcx>>, ) -> QueryResult<'tcx> { match crate::traits::wf::unnormalized_obligations( self.infcx, goal.param_env, goal.predicate, ) { Some(obligations) => self.evaluate_all_and_make_canonical_response( obligations.into_iter().map(|o| o.into()).collect(), ), None => self.make_canonical_response(Certainty::AMBIGUOUS), } } #[instrument(level = "debug", skip(self), ret)] fn compute_alias_eq_goal( &mut self, goal: Goal<'tcx, (ty::Term<'tcx>, ty::Term<'tcx>)>, ) -> QueryResult<'tcx> { let tcx = self.tcx(); let evaluate_normalizes_to = |ecx: &mut EvalCtxt<'_, 'tcx>, alias, other| { debug!("evaluate_normalizes_to(alias={:?}, other={:?})", alias, other); let r = ecx.probe(|ecx| { let (_, certainty) = ecx.evaluate_goal(goal.with( tcx, ty::Binder::dummy(ty::ProjectionPredicate { projection_ty: alias, term: other, }), ))?; ecx.make_canonical_response(certainty) }); debug!("evaluate_normalizes_to(..) -> {:?}", r); r }; if goal.predicate.0.is_infer() || goal.predicate.1.is_infer() { bug!( "`AliasEq` goal with an infer var on lhs or rhs which should have been instantiated" ); } match ( goal.predicate.0.to_alias_term_no_opaque(tcx), goal.predicate.1.to_alias_term_no_opaque(tcx), ) { (None, None) => bug!("`AliasEq` goal without an alias on either lhs or rhs"), (Some(alias), None) => evaluate_normalizes_to(self, alias, goal.predicate.1), (None, Some(alias)) => evaluate_normalizes_to(self, alias, goal.predicate.0), (Some(alias_lhs), Some(alias_rhs)) => { debug!("compute_alias_eq_goal: both sides are aliases"); let mut candidates = Vec::with_capacity(3); // Evaluate all 3 potential candidates for the alias' being equal candidates.push(evaluate_normalizes_to(self, alias_lhs, goal.predicate.1)); candidates.push(evaluate_normalizes_to(self, alias_rhs, goal.predicate.0)); candidates.push(self.probe(|this| { debug!("compute_alias_eq_goal: alias defids are equal, equating substs"); let nested_goals = this.eq(goal.param_env, alias_lhs, alias_rhs)?; this.evaluate_all_and_make_canonical_response(nested_goals) })); debug!(?candidates); self.try_merge_responses(candidates.into_iter()) } } } #[instrument(level = "debug", skip(self), ret)] fn compute_const_arg_has_type_goal( &mut self, goal: Goal<'tcx, (ty::Const<'tcx>, Ty<'tcx>)>, ) -> QueryResult<'tcx> { let (ct, ty) = goal.predicate; let nested_goals = self.eq(goal.param_env, ct.ty(), ty)?; self.evaluate_all_and_make_canonical_response(nested_goals) } } impl<'tcx> EvalCtxt<'_, 'tcx> { // Recursively evaluates a list of goals to completion, returning the certainty // of all of the goals. fn evaluate_all( &mut self, mut goals: Vec>>, ) -> Result { let mut new_goals = Vec::new(); self.repeat_while_none( |_| Ok(Certainty::Maybe(MaybeCause::Overflow)), |this| { let mut has_changed = Err(Certainty::Yes); for goal in goals.drain(..) { let (changed, certainty) = match this.evaluate_goal(goal) { Ok(result) => result, Err(NoSolution) => return Some(Err(NoSolution)), }; if changed { has_changed = Ok(()); } match certainty { Certainty::Yes => {} Certainty::Maybe(_) => { new_goals.push(goal); has_changed = has_changed.map_err(|c| c.unify_and(certainty)); } } } match has_changed { Ok(()) => { mem::swap(&mut new_goals, &mut goals); None } Err(certainty) => Some(Ok(certainty)), } }, ) } // Recursively evaluates a list of goals to completion, making a query response. // // This is just a convenient way of calling [`EvalCtxt::evaluate_all`], // then [`EvalCtxt::make_canonical_response`]. fn evaluate_all_and_make_canonical_response( &mut self, goals: Vec>>, ) -> QueryResult<'tcx> { self.evaluate_all(goals).and_then(|certainty| self.make_canonical_response(certainty)) } fn try_merge_responses( &mut self, responses: impl Iterator>, ) -> QueryResult<'tcx> { let candidates = responses.into_iter().flatten().collect::>(); if candidates.is_empty() { return Err(NoSolution); } // FIXME(-Ztreat-solver=next): We should instead try to find a `Certainty::Yes` response with // a subset of the constraints that all the other responses have. let one = candidates[0]; if candidates[1..].iter().all(|resp| resp == &one) { return Ok(one); } if let Some(response) = candidates.iter().find(|response| { response.value.certainty == Certainty::Yes && response.has_no_inference_or_external_constraints() }) { return Ok(*response); } let certainty = candidates.iter().fold(Certainty::AMBIGUOUS, |certainty, response| { certainty.unify_and(response.value.certainty) }); // FIXME(-Ztrait-solver=next): We should take the intersection of the constraints on all the // responses and use that for the constraints of this ambiguous response. let response = self.make_canonical_response(certainty); if let Ok(response) = &response { assert!(response.has_no_inference_or_external_constraints()); } response } } pub(super) fn response_no_constraints<'tcx>( tcx: TyCtxt<'tcx>, goal: Canonical<'tcx, impl Sized>, certainty: Certainty, ) -> QueryResult<'tcx> { Ok(Canonical { max_universe: goal.max_universe, variables: goal.variables, value: Response { var_values: CanonicalVarValues::make_identity(tcx, goal.variables), // FIXME: maybe we should store the "no response" version in tcx, like // we do for tcx.types and stuff. external_constraints: tcx.mk_external_constraints(ExternalConstraintsData::default()), certainty, }, }) }