Don't build `ParamEnv` and do trait solving in `ItemCtxt`s when lowering IATs #140247

BoxyUwU · 2025-04-24T10:15:25Z

Fixes #125879
This was due to updating inhabited predicate stuff which I had to do to make constructing ADTs with IATs in fields not ICE

Fixes #136678 (but no test added, I don't rly care about weird IAT edge cases under GCE)

Fixes #138240
Fixes #138266
These ICEs got masked but we have 3 other ICEs that look like dupes so I'm not too worried about this. Caused by mgca not properly lowering paths in type signatures to forbid infers or something, and now we happen to emit an error when we previously didnt.

Avoids doing "fully correct" candidate selection for IATs during hir ty lowering when in item signatures as it almost always leads to a query cycle from trying to build a ParamEnv. I replaced it with a use DeepRejectCtxt which should be able to handle this kind of conservative "could these types unify" while in a context where we don't want to do type equality.

A potential alternative would be to replace all the generic parameters in the self type with inference variables as then it would be "truly" environment agnostic and we could just use an empty environment. I think this is somewhat subpar as we would be weakening how much we used the self type to guide candidate selection and that's the only way for the user to attempt to specify which IAT to resolve to.

In general that's also why I think doing something hacky here instead of just throwing our hands up and saying "ItemCtxts shouldnt make inference variables" and not filtering any candidates out, is a good idea (there is no way for users to disambiguate such cases).

I'm not really sure how this interacts with #126651, though I'm also not really sure its super important to support projecting IATs from IAT self types given we don't even support T::Assoc::Other for trait-associated types so didn't give much thought to how this might fit in with that.

r? @compiler-errors
cc @fmease

rustbot · 2025-04-24T10:15:31Z

This PR changes a file inside tests/crashes. If a crash was fixed, please move into the corresponding ui subdir and add 'Fixes #' to the PR description to autoclose the issue upon merge.

HIR ty lowering was modified

cc @fmease

compiler/rustc_hir_analysis/src/collect.rs

bors · 2025-04-28T21:32:11Z

☔ The latest upstream changes (presumably #140388) made this pull request unmergeable. Please resolve the merge conflicts.

compiler-errors · 2025-05-22T09:49:10Z

@bors r+

bors · 2025-05-22T09:49:13Z

📌 Commit 34ce403 has been approved by compiler-errors

It is now in the queue for this repository.

…iaskrgr Rollup of 7 pull requests Successful merges: - rust-lang#135562 (Add ignore value suggestion in closure body) - rust-lang#139635 (Finalize repeat expr inference behaviour with inferred repeat counts) - rust-lang#139668 (Handle regions equivalent to 'static in non_local_bounds) - rust-lang#140218 (HIR ty lowering: Clean up & refactor the lowering of type-relative paths) - rust-lang#140435 (use uX::from instead of _ as uX in non - const contexts) - rust-lang#141130 (rustc_on_unimplemented cleanups) - rust-lang#141286 (Querify `coroutine_hidden_types`) Failed merges: - rust-lang#140247 (Don't build `ParamEnv` and do trait solving in `ItemCtxt`s when lowering IATs) r? `@ghost` `@rustbot` modify labels: rollup

bors · 2025-05-23T01:02:00Z

☔ The latest upstream changes (presumably #141396) made this pull request unmergeable. Please resolve the merge conflicts.

BoxyUwU · 2025-05-23T14:28:28Z

I should probably update this to use DeepRejectCtxt and handle LTAs correctly like discussed at the all hands 🤔 I was just gonna do it in a followup PR but if I'm having to rebase and figure out new crashes tests then I may aswell just do it all now

compiler/rustc_hir_analysis/src/collect.rs

BoxyUwU · 2025-05-23T16:23:00Z

compiler/rustc_hir_analysis/src/collect.rs

+        // We make an infcx and replace any escaping vars with placeholders so that IAT res
+        // with type/const bound vars in arguments is slightly smarter. `for<T> <Foo<T>>::IAT`
+        // would otherwise unify with `impl Foo<u8>` when ideally we would not.
+        let infcx = self.tcx().infer_ctxt().build(TypingMode::non_body_analysis());


Creating an infcx just to make placeholders feels pretty bad lol

do u need to? id expect expand free aliases to handle escaping bound vars. so does deep reject

deep reject "handles" escaping vars by allowing them to unify with anything, which is weaker IAT resolution than replacing with placeholders. See tests/ui/associated-inherent-types/bound_vars_in_args.rs which fails if we don't replace with placeholders

BoxyUwU · 2025-05-23T19:06:37Z

compiler/rustc_hir_analysis/src/hir_ty_lowering/mod.rs

-        if candidates.is_empty() {
-            return Ok(None);
-        }


Removing this masked a bunch of mgca ICEs I think. We now emit errors in all cases where IAT resolution fails to find any candidates which makes some delayed bugs no longer trigger. I've replaced a few of the crashes tests with examples that don't rely on IATs and I removed the two others that I couldn't find a reproducer for that didn't seem like a dupe of the other three.

…iaskrgr Rollup of 7 pull requests Successful merges: - rust-lang#135562 (Add ignore value suggestion in closure body) - rust-lang#139635 (Finalize repeat expr inference behaviour with inferred repeat counts) - rust-lang#139668 (Handle regions equivalent to 'static in non_local_bounds) - rust-lang#140218 (HIR ty lowering: Clean up & refactor the lowering of type-relative paths) - rust-lang#140435 (use uX::from instead of _ as uX in non - const contexts) - rust-lang#141130 (rustc_on_unimplemented cleanups) - rust-lang#141286 (Querify `coroutine_hidden_types`) Failed merges: - rust-lang#140247 (Don't build `ParamEnv` and do trait solving in `ItemCtxt`s when lowering IATs) r? `@ghost` `@rustbot` modify labels: rollup

BoxyUwU · 2025-05-27T17:28:50Z

tests/ui/associated-inherent-types/impl_params_are_infers.rs

+    // It would be illegal to resolve to `Foo<T, T, u8>::IAT`  as  `T` and `U` are
+    // different types. However, currently we treat all impl-side params sort of like
+    // they're infers and assume they can unify with anything, so we consider it a
+    // valid candidate.
+    field: Foo<T, U, <u16 as Identity>::This>::IAT,
+    //~^ ERROR: multiple applicable items in scope


Honestly not sure how I feel about using DeepRejectCtxt knowing it doesn't let this code work 🤔 This PR is already a big step forwards though so I think I'm in favour of just continuing with this impl and revisiting before stabilization.

bors · 2025-05-27T19:39:03Z

☔ The latest upstream changes (presumably #141644) made this pull request unmergeable. Please resolve the merge conflicts.

BoxyUwU · 2025-05-27T19:44:38Z

oh come on lol

lcnr · 2025-06-11T10:05:04Z

compiler/rustc_hir_typeck/src/fn_ctxt/mod.rs

+                            let ocx = ObligationCtxt::new_with_diagnostics(self);
+                            let self_ty =
+                                ocx.normalize(&ObligationCause::dummy(), self.param_env, self_ty);
+
+                            let impl_args = infcx.fresh_args_for_item(span, impl_);
+                            let impl_ty = tcx.type_of(impl_).instantiate(tcx, impl_args);
+                            let impl_ty =
+                                ocx.normalize(&ObligationCause::dummy(), self.param_env, impl_ty);
+
+                            // Check that the self types can be related.
+                            if ocx
+                                .eq(&ObligationCause::dummy(), self.param_env, impl_ty, self_ty)
+                                .is_err()
+                            {
+                                return false;
+                            }
+
+                            // Check whether the impl imposes obligations we have to worry about.
+                            let impl_bounds = tcx.predicates_of(impl_).instantiate(tcx, impl_args);
+                            let impl_bounds = ocx.normalize(
+                                &ObligationCause::dummy(),
+                                self.param_env,
+                                impl_bounds,
+                            );
+                            let impl_obligations = traits::predicates_for_generics(
+                                |_, _| ObligationCause::dummy(),
+                                self.param_env,
+                                impl_bounds,
+                            );
+                            ocx.register_obligations(impl_obligations);
+
+                            let mut errors = ocx.select_where_possible();
+                            if !errors.is_empty() {
+                                fulfillment_errors.append(&mut errors);
+                                return false;
+                            }
+
+                            true


move this into a sub_fn

lcnr · 2025-06-11T10:07:57Z

compiler/rustc_type_ir/src/fast_reject.rs

+        if lhs == rhs {
+            return true;
+        }


this fast path should be in types_may_unify_inner

rustbot assigned compiler-errors Apr 24, 2025

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Apr 24, 2025

BoxyUwU commented Apr 24, 2025

View reviewed changes

compiler/rustc_hir_analysis/src/collect.rs Outdated Show resolved Hide resolved

compiler-errors reviewed Apr 24, 2025

View reviewed changes

compiler/rustc_hir_analysis/src/collect.rs Outdated Show resolved Hide resolved

BoxyUwU force-pushed the iats_no_query_cycles branch from 3361c7d to 34ce403 Compare April 29, 2025 15:06

fmease mentioned this pull request Apr 29, 2025

Tracking Issue for RFC 3591: useing a function from a trait #134691

Open

4 tasks

compiler-errors approved these changes May 22, 2025

View reviewed changes

bors 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-review Status: Awaiting review from the assignee but also interested parties. labels May 22, 2025

matthiaskrgr mentioned this pull request May 22, 2025

Rollup of 7 pull requests #141396

Merged

bors 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-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels May 23, 2025

BoxyUwU force-pushed the iats_no_query_cycles branch 2 times, most recently from aee1a22 to 701d3b8 Compare May 23, 2025 14:25

BoxyUwU force-pushed the iats_no_query_cycles branch from 701d3b8 to 276daf2 Compare May 23, 2025 15:16

BoxyUwU commented May 23, 2025

View reviewed changes

compiler/rustc_hir_analysis/src/collect.rs Outdated Show resolved Hide resolved

lcnr reviewed May 23, 2025

View reviewed changes

compiler/rustc_hir_analysis/src/collect.rs Outdated Show resolved Hide resolved

BoxyUwU force-pushed the iats_no_query_cycles branch from 276daf2 to 5867bb5 Compare May 23, 2025 16:22

BoxyUwU commented May 23, 2025

View reviewed changes

This comment has been minimized.

Sign in to view

BoxyUwU force-pushed the iats_no_query_cycles branch from 5867bb5 to 02b87a8 Compare May 23, 2025 18:06

BoxyUwU commented May 23, 2025

View reviewed changes

BoxyUwU force-pushed the iats_no_query_cycles branch from 02b87a8 to ccec360 Compare May 23, 2025 19:16

BoxyUwU 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 May 23, 2025

BoxyUwU force-pushed the iats_no_query_cycles branch from 7fae1aa to 8f33b8f Compare May 27, 2025 17:25

BoxyUwU commented May 27, 2025

View reviewed changes

BoxyUwU added 2 commits May 29, 2025 19:00

Don't build ParamEnv and do trait solving in ItemCtxts

3f04713

Test whether we use DeepRejectCtxt

8f2995c

BoxyUwU force-pushed the iats_no_query_cycles branch from 8f33b8f to 8f2995c Compare May 29, 2025 18:00

fmease added the F-inherent_associated_types `#![feature(inherent_associated_types)]` label Jun 3, 2025

fmease added this to Inherent Associated Types (IAT) Jun 3, 2025

fmease moved this to In Progress in Inherent Associated Types (IAT) Jun 3, 2025

fmease mentioned this pull request Jun 4, 2025

If inherent associated type selection fails, we should still consider trait associated type candidates as a fallback #142006

Open

lcnr reviewed Jun 11, 2025

View reviewed changes

BoxyUwU assigned lcnr and unassigned compiler-errors Jun 11, 2025

Don't build ParamEnv and do trait solving in ItemCtxts when lowering IATs #140247

Are you sure you want to change the base?

Don't build ParamEnv and do trait solving in ItemCtxts when lowering IATs #140247

Uh oh!

Conversation

BoxyUwU commented Apr 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Apr 24, 2025

Uh oh!

Uh oh!

Uh oh!

bors commented Apr 28, 2025

Uh oh!

compiler-errors commented May 22, 2025

Uh oh!

bors commented May 22, 2025

Uh oh!

bors commented May 23, 2025

Uh oh!

BoxyUwU commented May 23, 2025

Uh oh!

Uh oh!

Uh oh!

BoxyUwU May 23, 2025

Choose a reason for hiding this comment

Uh oh!

lcnr May 24, 2025

Choose a reason for hiding this comment

Uh oh!

BoxyUwU May 24, 2025 • 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!

This comment has been minimized.

BoxyUwU May 23, 2025

Choose a reason for hiding this comment

Uh oh!

BoxyUwU May 27, 2025

Choose a reason for hiding this comment

Uh oh!

bors commented May 27, 2025

Uh oh!

BoxyUwU commented May 27, 2025

Uh oh!

lcnr Jun 11, 2025

Choose a reason for hiding this comment

Uh oh!

lcnr Jun 11, 2025

Choose a reason for hiding this comment

Uh oh!

lcnr Jun 11, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Don't build `ParamEnv` and do trait solving in `ItemCtxt`s when lowering IATs #140247

Don't build `ParamEnv` and do trait solving in `ItemCtxt`s when lowering IATs #140247

BoxyUwU commented Apr 24, 2025 •

edited

Loading

BoxyUwU May 24, 2025 •

edited

Loading