trait_selection: Keep type-op region constraints in borrowck - #161423
trait_selection: Keep type-op region constraints in borrowck#161423Dnreikronos wants to merge 2 commits into
Conversation
Keep rust-lang#158588 focused on reporting solver region constraints. The type-op behavior and its borrowck coverage now live in rust-lang#161423.
|
r? @davidtwco rustbot has assigned @davidtwco. Use Why was this reviewer chosen?The reviewer was selected based on:
|
|
r? BoxyUwU |
|
I think this is a partial revert of #160982 which we don't want. I think that teaching @rustbot author |
|
Reminder, once the PR becomes ready for a review, use |
…tions_binders_diagnostics, r=BoxyUwU trait_selection: fix assumptions-on-binders diagnostics fixes rust-lang#157732 `-Zassumptions-on-binders` was losing the origin of solver region constraints. By the time regionck or borrowck reported them, every constraint received the containing item span, so diagnostics pointed at an entire function or const. Some paths were also still emitting placeholder text (`:3` and `meoow :c`). This keeps canonical solver responses and `ExternalConstraintsData` span-free, so source locations do not participate in candidate equality or caching. When a response is applied, `EvalCtxt::origin_span` is attached to each atomic solver `RegionConstraint` stored in `InferCtxt`. Late region conversion then uses the span attached to each constraint, including for ambiguity diagnostics. The affected paths now report `higher-ranked lifetime bound could not be satisfied` at the type use that introduced the failing constraint. UI coverage checks the rust-lang#157732 call-site diagnostic and the existing regionck alias-outlives case. Unit coverage checks that the spanned evaluator stays in semantic parity with the type-ir evaluator and preserves the first ambiguity origin span. The type-op behavior I ran into while working on this is split into rust-lang#161423. Future work: carry the failed outlives predicate or originating binder far enough through this path to name the exact bound that failed, rather than only pointing at its origin.
…tions_binders_diagnostics, r=BoxyUwU trait_selection: fix assumptions-on-binders diagnostics fixes rust-lang#157732 `-Zassumptions-on-binders` was losing the origin of solver region constraints. By the time regionck or borrowck reported them, every constraint received the containing item span, so diagnostics pointed at an entire function or const. Some paths were also still emitting placeholder text (`:3` and `meoow :c`). This keeps canonical solver responses and `ExternalConstraintsData` span-free, so source locations do not participate in candidate equality or caching. When a response is applied, `EvalCtxt::origin_span` is attached to each atomic solver `RegionConstraint` stored in `InferCtxt`. Late region conversion then uses the span attached to each constraint, including for ambiguity diagnostics. The affected paths now report `higher-ranked lifetime bound could not be satisfied` at the type use that introduced the failing constraint. UI coverage checks the rust-lang#157732 call-site diagnostic and the existing regionck alias-outlives case. Unit coverage checks that the spanned evaluator stays in semantic parity with the type-ir evaluator and preserves the first ambiguity origin span. The type-op behavior I ran into while working on this is split into rust-lang#161423. Future work: carry the failed outlives predicate or originating binder far enough through this path to name the exact bound that failed, rather than only pointing at its origin.
…tions_binders_diagnostics, r=BoxyUwU trait_selection: fix assumptions-on-binders diagnostics fixes rust-lang#157732 `-Zassumptions-on-binders` was losing the origin of solver region constraints. By the time regionck or borrowck reported them, every constraint received the containing item span, so diagnostics pointed at an entire function or const. Some paths were also still emitting placeholder text (`:3` and `meoow :c`). This keeps canonical solver responses and `ExternalConstraintsData` span-free, so source locations do not participate in candidate equality or caching. When a response is applied, `EvalCtxt::origin_span` is attached to each atomic solver `RegionConstraint` stored in `InferCtxt`. Late region conversion then uses the span attached to each constraint, including for ambiguity diagnostics. The affected paths now report `higher-ranked lifetime bound could not be satisfied` at the type use that introduced the failing constraint. UI coverage checks the rust-lang#157732 call-site diagnostic and the existing regionck alias-outlives case. Unit coverage checks that the spanned evaluator stays in semantic parity with the type-ir evaluator and preserves the first ambiguity origin span. The type-op behavior I ran into while working on this is split into rust-lang#161423. Future work: carry the failed outlives predicate or originating binder far enough through this path to name the exact bound that failed, rather than only pointing at its origin.
…tions_binders_diagnostics, r=BoxyUwU trait_selection: fix assumptions-on-binders diagnostics fixes rust-lang#157732 `-Zassumptions-on-binders` was losing the origin of solver region constraints. By the time regionck or borrowck reported them, every constraint received the containing item span, so diagnostics pointed at an entire function or const. Some paths were also still emitting placeholder text (`:3` and `meoow :c`). This keeps canonical solver responses and `ExternalConstraintsData` span-free, so source locations do not participate in candidate equality or caching. When a response is applied, `EvalCtxt::origin_span` is attached to each atomic solver `RegionConstraint` stored in `InferCtxt`. Late region conversion then uses the span attached to each constraint, including for ambiguity diagnostics. The affected paths now report `higher-ranked lifetime bound could not be satisfied` at the type use that introduced the failing constraint. UI coverage checks the rust-lang#157732 call-site diagnostic and the existing regionck alias-outlives case. Unit coverage checks that the spanned evaluator stays in semantic parity with the type-ir evaluator and preserves the first ambiguity origin span. The type-op behavior I ran into while working on this is split into rust-lang#161423. Future work: carry the failed outlives predicate or originating binder far enough through this path to name the exact bound that failed, rather than only pointing at its origin.
…tions_binders_diagnostics, r=BoxyUwU trait_selection: fix assumptions-on-binders diagnostics fixes rust-lang#157732 `-Zassumptions-on-binders` was losing the origin of solver region constraints. By the time regionck or borrowck reported them, every constraint received the containing item span, so diagnostics pointed at an entire function or const. Some paths were also still emitting placeholder text (`:3` and `meoow :c`). This keeps canonical solver responses and `ExternalConstraintsData` span-free, so source locations do not participate in candidate equality or caching. When a response is applied, `EvalCtxt::origin_span` is attached to each atomic solver `RegionConstraint` stored in `InferCtxt`. Late region conversion then uses the span attached to each constraint, including for ambiguity diagnostics. The affected paths now report `higher-ranked lifetime bound could not be satisfied` at the type use that introduced the failing constraint. UI coverage checks the rust-lang#157732 call-site diagnostic and the existing regionck alias-outlives case. Unit coverage checks that the spanned evaluator stays in semantic parity with the type-ir evaluator and preserves the first ambiguity origin span. The type-op behavior I ran into while working on this is split into rust-lang#161423. Future work: carry the failed outlives predicate or originating binder far enough through this path to name the exact bound that failed, rather than only pointing at its origin.
…tions_binders_diagnostics, r=BoxyUwU trait_selection: fix assumptions-on-binders diagnostics fixes rust-lang#157732 `-Zassumptions-on-binders` was losing the origin of solver region constraints. By the time regionck or borrowck reported them, every constraint received the containing item span, so diagnostics pointed at an entire function or const. Some paths were also still emitting placeholder text (`:3` and `meoow :c`). This keeps canonical solver responses and `ExternalConstraintsData` span-free, so source locations do not participate in candidate equality or caching. When a response is applied, `EvalCtxt::origin_span` is attached to each atomic solver `RegionConstraint` stored in `InferCtxt`. Late region conversion then uses the span attached to each constraint, including for ambiguity diagnostics. The affected paths now report `higher-ranked lifetime bound could not be satisfied` at the type use that introduced the failing constraint. UI coverage checks the rust-lang#157732 call-site diagnostic and the existing regionck alias-outlives case. Unit coverage checks that the spanned evaluator stays in semantic parity with the type-ir evaluator and preserves the first ambiguity origin span. The type-op behavior I ran into while working on this is split into rust-lang#161423. Future work: carry the failed outlives predicate or originating binder far enough through this path to name the exact bound that failed, rather than only pointing at its origin.
…tions_binders_diagnostics, r=BoxyUwU trait_selection: fix assumptions-on-binders diagnostics fixes rust-lang#157732 `-Zassumptions-on-binders` was losing the origin of solver region constraints. By the time regionck or borrowck reported them, every constraint received the containing item span, so diagnostics pointed at an entire function or const. Some paths were also still emitting placeholder text (`:3` and `meoow :c`). This keeps canonical solver responses and `ExternalConstraintsData` span-free, so source locations do not participate in candidate equality or caching. When a response is applied, `EvalCtxt::origin_span` is attached to each atomic solver `RegionConstraint` stored in `InferCtxt`. Late region conversion then uses the span attached to each constraint, including for ambiguity diagnostics. The affected paths now report `higher-ranked lifetime bound could not be satisfied` at the type use that introduced the failing constraint. UI coverage checks the rust-lang#157732 call-site diagnostic and the existing regionck alias-outlives case. Unit coverage checks that the spanned evaluator stays in semantic parity with the type-ir evaluator and preserves the first ambiguity origin span. The type-op behavior I ran into while working on this is split into rust-lang#161423. Future work: carry the failed outlives predicate or originating binder far enough through this path to name the exact bound that failed, rather than only pointing at its origin.
…tions_binders_diagnostics, r=BoxyUwU trait_selection: fix assumptions-on-binders diagnostics fixes rust-lang#157732 `-Zassumptions-on-binders` was losing the origin of solver region constraints. By the time regionck or borrowck reported them, every constraint received the containing item span, so diagnostics pointed at an entire function or const. Some paths were also still emitting placeholder text (`:3` and `meoow :c`). This keeps canonical solver responses and `ExternalConstraintsData` span-free, so source locations do not participate in candidate equality or caching. When a response is applied, `EvalCtxt::origin_span` is attached to each atomic solver `RegionConstraint` stored in `InferCtxt`. Late region conversion then uses the span attached to each constraint, including for ambiguity diagnostics. The affected paths now report `higher-ranked lifetime bound could not be satisfied` at the type use that introduced the failing constraint. UI coverage checks the rust-lang#157732 call-site diagnostic and the existing regionck alias-outlives case. Unit coverage checks that the spanned evaluator stays in semantic parity with the type-ir evaluator and preserves the first ambiguity origin span. The type-op behavior I ran into while working on this is split into rust-lang#161423. Future work: carry the failed outlives predicate or originating binder far enough through this path to name the exact bound that failed, rather than only pointing at its origin.
Rollup merge of #158588 - Dnreikronos:trait_selection/assumptions_binders_diagnostics, r=BoxyUwU trait_selection: fix assumptions-on-binders diagnostics fixes #157732 `-Zassumptions-on-binders` was losing the origin of solver region constraints. By the time regionck or borrowck reported them, every constraint received the containing item span, so diagnostics pointed at an entire function or const. Some paths were also still emitting placeholder text (`:3` and `meoow :c`). This keeps canonical solver responses and `ExternalConstraintsData` span-free, so source locations do not participate in candidate equality or caching. When a response is applied, `EvalCtxt::origin_span` is attached to each atomic solver `RegionConstraint` stored in `InferCtxt`. Late region conversion then uses the span attached to each constraint, including for ambiguity diagnostics. The affected paths now report `higher-ranked lifetime bound could not be satisfied` at the type use that introduced the failing constraint. UI coverage checks the #157732 call-site diagnostic and the existing regionck alias-outlives case. Unit coverage checks that the spanned evaluator stays in semantic parity with the type-ir evaluator and preserves the first ambiguity origin span. The type-op behavior I ran into while working on this is split into #161423. Future work: carry the failed outlives predicate or originating binder far enough through this path to name the exact bound that failed, rather than only pointing at its origin.
65f9dd2 to
2959d13
Compare
|
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. |
|
Sup,@BoxyUwU :) Yeah, The next solver was storing the tree in the temporary query I also found another place dropping the same tree during implied-bound normalization, before lexical regionck. I fixed that too and added regressions for that path and the borrowck case. I like this version a lot more. It fixes the point where the data was lost and keeps the cache behavior from #160982. The local version worked, but mostly by going around the problem, and that felt wrong once I understood the full path. @rustbot ready |
I ran into this while working on #158588.
borrowck_env_failstill had a FIXME because the function body wasn't reporting the outlives error. The next solver creates the region constraint, but the canonical type-op path doesn't put it inQueryResponse. Borrowck never sees it, so the type op looks fine and the error is lost.With
-Zassumptions-on-binders, these type ops now run locally on borrowck'sInferCtxt. The fast path still runs first. That leaves the constraint in the same inference context borrowck reads later.I prefer this over adding more data to the old canonical response. The next solver already caches its work, and teaching the old query path about these constraints felt like extra machinery for something we can avoid. Running locally is pretty boring, but I think that's a good thing here. The old FIXME is now the regression test.