trait_selection: Keep type-op region constraints in borrowck by Dnreikronos · Pull Request #161423 · rust-lang/rust

Dnreikronos · 2026-08-20T19:20:19Z

I ran into this while working on #158588. borrowck_env_fail still 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 in QueryResponse. 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's InferCtxt. 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.

Keep rust-lang#158588 focused on reporting solver region constraints. The type-op behavior and its borrowck coverage now live in rust-lang#161423.

rustbot · 2026-08-20T19:33:12Z

r? @davidtwco

rustbot has assigned @davidtwco.
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

Why was this reviewer chosen?

The reviewer was selected based on:

Owners of files modified in this PR: compiler, types
compiler, types expanded to 75 candidates
Random selection from 17 candidates

BoxyUwU · 2026-08-20T21:13:43Z

r? BoxyUwU

BoxyUwU · 2026-08-21T19:39:46Z

I think this is a partial revert of #160982 which we don't want. I think that teaching scrape_region_constraints to handle the new kind of constraints is exactly correct. The whole point of that function is to return the region constraints so it not doing that under -Zassumptions-on-binders seems like a clear bug to me :3

@rustbot author

rustbot · 2026-08-21T19:39:49Z

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

…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.

Dnreikronos · 2026-08-22T23:47:48Z

Sup,@BoxyUwU :)

Yeah, scrape_region_constraints makes more sense here. I went with the local type op first because it made the new constraints survive, but after tracing it a bit more I realized, as you said, that was just dodging the query path from #160982.

The next solver was storing the tree in the temporary query InferCtxt, but the query result only knew about the old region constraints. So when that context went away, the tree went with it. I changed QueryRegionConstraints to carry the tree too. It stays unspanned in the cached result, and the caller adds the local span when it registers it. scrape_region_constraints now returns the constraints made by that op without messing with whatever the caller already had.

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

rustbot · 2026-08-29T14:06:54Z

Some changes occurred to the core trait solver

cc @rust-lang/initiative-trait-system-refactor

Dnreikronos · 2026-08-29T14:12:34Z

@rustbot ready

rustbot · 2026-09-04T14:07:44Z

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.

BoxyUwU

this looks pretty much done now :)

View changes since this review

BoxyUwU · 2026-09-08T12:14:18Z

        type_tests,
+        // These have already been destructured into `outlives_constraints` at the
+        // end of MIR type checking.
+        solver_constraints: _,


want to add an assert that this is empty? that way the comment cant go out of date

yep....the comment is an assert now. It binds the field and checks it's still trivially true right after the destructure, so if something starts pushing constraints in there after typeck we get a panic instead of quietly dropping them on the floor.

I almost left it as a comment because I'd already convinced myself nothing could get in there. But that's exactly the sort of thing I'd forget in two months, so.

BoxyUwU · 2026-09-08T12:35:23Z

                region_obligations,
                region_constraints,
                region_assumptions,
+                Default::default(),


Can you add a comment about why we ignore new style constraints here

Added. Short version is this only runs with the flag off, and with the flag off the solver never makes new style constraints in the first place. Turn it on and we go down the NextGen path instead, which reads the tree straight off the InferCtxt, so this function doesn't even get called.

Took me a minute to talk myself into that when I first passed the empty default in, so the comment probably saves the next person the same trip.

BoxyUwU · 2026-09-08T12:37:49Z

+/// Executes `op` and then scrapes out all resulting region constraints,
+/// creating query-region-constraints.


Suggested change

/// Executes `op` and then scrapes out all resulting region constraints,

/// creating query-region-constraints.

/// Executes `op` and then scrapes out all resulting region constraints

/// in the `infcx`, creating query-region-constraints.

Applied. "in the infcx" is the part that actually carries the meaning, my version was vaguer for no reason.

BoxyUwU · 2026-09-08T12:39:13Z

            region_obligations,
            &region_constraints,
            region_assumptions,
+            Default::default(),


comment about why we dont care about new style constraints here

Added. This one returns early unless we're on the old solver globally, and the flag switches the new solver on globally, so the two can't ever be true at the same time. There's just never anything to pass along here.

BoxyUwU · 2026-09-08T12:39:50Z

        let region_obligations = self.take_registered_region_obligations();
        let region_assumptions = self.take_registered_region_assumptions();
        debug!(?region_obligations);
+        let solver_constraints = self.clone_solver_region_constraints();


why not take_solver_region_constraints like we do for the other stuff?

No good reason...that's leftover from the version with the swap, where the tree had to stay in the InferCtxt because borrowck read it much later, so cloning was the only option. Once the swap went away the clone was pure habit and I didn't notice.

Switched to take. We already take the region obligations and assumptions a couple of lines above, so we can't be in a snapshot there and the assert inside take is fine. clone_solver_region_constraints had no other callers, so it's deleted.

BoxyUwU · 2026-09-08T12:40:21Z

want to destructure query_response.value.region_constraints into all of its fields by pattern matching so that its clear we're not forgetting anything

Done. I did the same in the NLL variant since it reads the same three fields and would go stale the same way.

It also just reads better. Three separate walks down query_response.value.region_constraints and I was counting fields by hand to be sure I hadn't skipped one, which is the thing the pattern match is there to stop.

BoxyUwU · 2026-09-08T12:43:23Z

@rustbot author

BoxyUwU

great yeah this looks good, can you squash your commits then I'll approve :)

View changes since this review

Under `-Zassumptions-on-binders` the next solver stores its region constraints in the `InferCtxt` instead of registering region obligations. The canonical type-op path never copied them into `QueryResponse`, so they were dropped once the query's inference context went away. Borrowck then saw a type op with no constraints and lost the outlives error entirely. `QueryRegionConstraints` now carries the constraint next to the old style constraints and assumptions. It stays unspanned while passing through a canonical query and the caller attaches its own origin span when consuming the response. Borrowck accumulates these in `MirTypeckRegionConstraints` through `ConstraintConversion`, the same way it handles everything else, and destructures them into NLL outlives constraints at the end of MIR type checking. Implied bound normalization was dropping the same constraint before lexical regionck, so that path registers it now as well.

Dnreikronos · 2026-09-08T14:38:12Z

great yeah this looks good, can you squash your commits then I'll approve :)

View changes since this review

Did, Boxy :)

BoxyUwU · 2026-09-08T15:31:23Z

@bors r+

rust-bors · 2026-09-08T15:31:26Z

📌 Commit a80f20d has been approved by BoxyUwU

It is now in the queue for this repository.

🌲 The tree is currently closed for pull requests below priority 5. This pull request will be tested once the tree is reopened.

Reason for tree closure: spurious failures

…ve_type_op_constraints, r=BoxyUwU trait_selection: Keep type-op region constraints in borrowck I ran into this while working on rust-lang#158588. `borrowck_env_fail` still 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 in `QueryResponse`. 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's `InferCtxt`. 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.

…uwer Rollup of 6 pull requests Successful merges: - #162309 (offload: automate manual clang-linker-wrapper step) - #160505 (delegation: supporting inherent impls) - #160712 (windows-gnullvm: always link libunwind statically) - #161423 (trait_selection: Keep type-op region constraints in borrowck) - #162461 (limit the api of `fold_predicate` and `visit_predicate`) - #162475 (Fix unsoundness bug on next trait solver for dyn const generics placeholder)

Rollup merge of #161423 - Dnreikronos:trait_selection/preserve_type_op_constraints, r=BoxyUwU trait_selection: Keep type-op region constraints in borrowck I ran into this while working on #158588. `borrowck_env_fail` still 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 in `QueryResponse`. 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's `InferCtxt`. 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.

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Aug 20, 2026

Dnreikronos mentioned this pull request Aug 20, 2026

trait_selection: fix assumptions-on-binders diagnostics #158588

Merged

Dnreikronos marked this pull request as ready for review August 20, 2026 19:33

rustbot added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Aug 20, 2026

rustbot assigned davidtwco Aug 20, 2026

rustbot removed the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Aug 20, 2026

rustbot assigned BoxyUwU and unassigned davidtwco Aug 20, 2026

rustbot removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Aug 21, 2026

rustbot added the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Aug 21, 2026

Dnreikronos force-pushed the trait_selection/preserve_type_op_constraints branch from 65f9dd2 to 2959d13 Compare August 22, 2026 23:42

This comment has been minimized.

Sign in to view

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Aug 22, 2026

rustbot added the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Aug 27, 2026

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Aug 29, 2026

BoxyUwU mentioned this pull request Sep 3, 2026

Assumptions on Binders rust-lang/goals#621

Open

Dnreikronos mentioned this pull request Sep 3, 2026

trait solver: Include implied outlives assumptions #162238

Open

This comment has been minimized.

Sign in to view

Dnreikronos force-pushed the trait_selection/preserve_type_op_constraints branch from 96e94dd to 0f14fc7 Compare September 4, 2026 14:07

rustbot added the WG-trait-system-refactor The Rustc Trait System Refactor Initiative (-Znext-solver) label Sep 4, 2026

BoxyUwU reviewed Sep 8, 2026

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 Sep 8, 2026

BoxyUwU approved these changes Sep 8, 2026

View reviewed changes

Dnreikronos force-pushed the trait_selection/preserve_type_op_constraints branch from 9c65b1e to a80f20d Compare September 8, 2026 14:29

rust-bors Bot added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Sep 8, 2026

This was referenced Sep 8, 2026

Rollup of 7 pull requests #162486

Closed

Rollup of 6 pull requests #162487

Merged

rust-bors Bot merged commit 9318843 into rust-lang:main Sep 8, 2026
13 checks passed

rustbot added this to the 1.100.0 milestone Sep 8, 2026

		/// Executes `op` and then scrapes out all resulting region constraints,
		/// creating query-region-constraints.

Uh oh!

Uh oh!

Conversation

Dnreikronos commented Aug 20, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Aug 20, 2026

Uh oh!

BoxyUwU commented Aug 20, 2026

Uh oh!

BoxyUwU commented Aug 21, 2026

Uh oh!

rustbot commented Aug 21, 2026

Uh oh!

This comment has been minimized.

Dnreikronos commented Aug 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Aug 29, 2026

Uh oh!

Dnreikronos commented Aug 29, 2026

Uh oh!

This comment has been minimized.

rustbot commented Sep 4, 2026

Uh oh!

BoxyUwU left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Dnreikronos Sep 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Dnreikronos Sep 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

BoxyUwU commented Sep 8, 2026

Uh oh!

BoxyUwU left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Dnreikronos commented Sep 8, 2026

Uh oh!

BoxyUwU commented Sep 8, 2026

Uh oh!

rust-bors Bot commented Sep 8, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Dnreikronos commented Aug 20, 2026 •

edited by rustbot

Loading

Dnreikronos commented Aug 22, 2026 •

edited

Loading

BoxyUwU left a comment •

edited by rustbot

Loading

Dnreikronos Sep 8, 2026 •

edited

Loading

Dnreikronos Sep 8, 2026 •

edited

Loading

BoxyUwU left a comment •

edited by rustbot

Loading