trait solver: Resolve region vars in max universe by Dnreikronos · Pull Request #157921 · rust-lang/rust

Dnreikronos · 2026-06-15T15:02:04Z

The ICE comes from computing the max universe of a region constraint after a region var has already been unified with something else. In that state, asking universe_of_lt directly can hit None and panic.

This makes the region visitor opportunistically resolve ReVars first, then inspect the resolved region. Type and const infer vars stay as-is because this path already bails on non-region inference before it needs more structure from them.

The `MaxUniverse` region visitor matched on `ReVar` terms and unwrapped `universe_of_lt`, which returns `None` for a variable that has already been unified with another region. Such resolved variables can reach the visitor while computing region assumptions under `-Zassumptions-on-binders`, causing an ICE. Resolve the variable first and inspect whatever it points at instead of assuming it is still an unresolved inference variable.

rustbot · 2026-06-15T15:02:10Z

r? @wesleywiser

rustbot has assigned @wesleywiser.
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
compiler expanded to 73 candidates
Random selection from 21 candidates

BoxyUwU · 2026-06-15T17:05:11Z

r? BoxyUwU

BoxyUwU

thx for looking into this :3

View changes since this review

BoxyUwU · 2026-06-19T13:49:37Z

+
+fn check/*<T>*/()
+where
+    [(); Parent::CT::<T>]:,


does adding the generic parameter T to check and introducing the lifetime 'a to check (and using it) cause this test to stop testing what it should? also is it possible to get a test case that doesn't depend on inherent_associated_types, the combo of mGCA and IATs is fairly brittle and likely to change significantly in the short future which will probably make this test stop testing anything 😅

BoxyUwU · 2026-06-19T13:53:58Z

            TyKind::Placeholder(p) => self.max_universe = self.max_universe.max(p.universe),
            TyKind::Infer(InferTy::TyVar(inf)) => {
+                // Unlike `visit_region`, we don't resolve the variable first: callers
+                // computing assumptions bail on any non-region inference variable


we use MaxUniverse when rewriting constraints to a lower universe which can happen on different terms than what we compute assumptions for (though there is significant overlap) 🤔 I would expect there to possibly be cases where we have unresolved type/const infer vars but im not sure.

I would honestly probably just put a resolve_vars_if_possible call in the max_universe function just to be safe and then say we do that here, instead of trying to rely on some kind of global reasoning for why stuff must be resolved

Dnreikronos · 2026-06-20T18:01:42Z

thanks, that makes sense.

imo putting resolve_vars_if_possible in max_universe is the better fix here, so i'll change that instead of relying on the assumptions caller having already ruled stuff out.

for the test, yeah, i don't love the iat + mgca dependency either. i'll try to reduce it to something that doesn't use iats. idk yet if the same resolved-region-var shape shows up without them, but i'll check before updating the test. ltm this is the part most likely to get brittle later.

Dnreikronos · 2026-06-20T19:27:51Z

i couldn't get a non-iat version to hit the same bug.

the closest things i tried were trait associated consts, with both explicit and inferred lifetimes, and a gat-ish associated type. they all just reported the expected user error and never reached the old max_universe ice, so i kept the iat/mgca test for now.

imo that's still better than landing a cleaner-looking test that doesn't cover the panic. idk if there is another way to force the same resolved-region-var constraint without iat, but i don't have one atm.

also, i moved resolve_vars_if_possible into max_universe and kept the explicit lifetime-var resolution, since that part is still needed for this ice path. lgtm to me as the less fragile fix, even if the test is still a bit annoying.

BoxyUwU · 2026-06-23T10:02:10Z

cool thanks, yeah if you can't find a test case without IATs then this is fine as is :)

@bors r+

rust-bors · 2026-06-23T10:02:13Z

📌 Commit a44ab8d has been approved by BoxyUwU

It is now in the queue for this repository.

…ders-placeholder-ice, r=BoxyUwU trait solver: Resolve region vars in max universe Fixes rust-lang#157862 The ICE comes from computing the max universe of a region constraint after a region var has already been unified with something else. In that state, asking universe_of_lt directly can hit None and panic. This makes the region visitor opportunistically resolve ReVars first, then inspect the resolved region. Type and const infer vars stay as-is because this path already bails on non-region inference before it needs more structure from them.

…uwer Rollup of 16 pull requests Successful merges: - #156885 (Fix misattributed type inference error span for index expressions) - #157271 (simplify some `proc_macro` things) - #157883 (Remove strict invariant node_type on hir_type during ty privacy visit) - #157921 (trait solver: Resolve region vars in max universe) - #157960 (delegation: add support for infers in generics) - #158105 (Extract all instance shim variants into new `ShimKind` enum) - #158207 (Resolver: local/external split of `resolve_ident_in_module_non_globs_unadjusted` ) - #158279 (Follow goto and drop when linting unreachable code) - #157807 (don't ice on non-lifetime binders under `-Zassumptions-on-binders`) - #158020 (Update mingw-w64 C toolchain) - #158222 (format: ignore println newline in foreign format hints) - #158223 (Move target checking for #[lang] to the attribute parser) - #158252 (Use `cfg_select` in `std::os`) - #158257 ( fix escaping placeholder check in next solver normalization folder) - #158274 (triagebot: Stop pinging myself) - #158282 (slice_split_once: bounds check optimization note) Failed merges: - #158256 (Avoid parser panics bubbling out to proc macros)

…ders-placeholder-ice, r=BoxyUwU trait solver: Resolve region vars in max universe Fixes rust-lang#157862 The ICE comes from computing the max universe of a region constraint after a region var has already been unified with something else. In that state, asking universe_of_lt directly can hit None and panic. This makes the region visitor opportunistically resolve ReVars first, then inspect the resolved region. Type and const infer vars stay as-is because this path already bails on non-region inference before it needs more structure from them.

…uwer Rollup of 23 pull requests Successful merges: - #158315 (`rust-analyzer` subtree update) - #155739 (Add temporary scope to assert_eq and assert_ne) - #156885 (Fix misattributed type inference error span for index expressions) - #157271 (simplify some `proc_macro` things) - #157883 (Remove strict invariant node_type on hir_type during ty privacy visit) - #157921 (trait solver: Resolve region vars in max universe) - #157960 (delegation: add support for infers in generics) - #157983 (Lift the same-signature restriction for `extern "tail"`) - #158105 (Extract all instance shim variants into new `ShimKind` enum) - #158207 (Resolver: local/external split of `resolve_ident_in_module_non_globs_unadjusted` ) - #158279 (Follow goto and drop when linting unreachable code) - #157527 (Move derive tests into their dedicated folder) - #157807 (don't ice on non-lifetime binders under `-Zassumptions-on-binders`) - #158020 (Update mingw-w64 C toolchain) - #158222 (format: ignore println newline in foreign format hints) - #158223 (Move target checking for #[lang] to the attribute parser) - #158252 (Use `cfg_select` in `std::os`) - #158257 ( fix escaping placeholder check in next solver normalization folder) - #158263 (Only load the feature list once in the entire resolver) - #158274 (triagebot: Stop pinging myself) - #158282 (slice_split_once: bounds check optimization note) - #158300 (Improve unknown crate_type diagnostic suggestions) - #158304 (mailmap: update mu001999) Failed merges: - #158256 (Avoid parser panics bubbling out to proc macros)

…ders-placeholder-ice, r=BoxyUwU trait solver: Resolve region vars in max universe Fixes rust-lang#157862 The ICE comes from computing the max universe of a region constraint after a region var has already been unified with something else. In that state, asking universe_of_lt directly can hit None and panic. This makes the region visitor opportunistically resolve ReVars first, then inspect the resolved region. Type and const infer vars stay as-is because this path already bails on non-region inference before it needs more structure from them.

Rollup of 27 pull requests Successful merges: - #158315 (`rust-analyzer` subtree update) - #155739 (Add temporary scope to assert_eq and assert_ne) - #156885 (Fix misattributed type inference error span for index expressions) - #157271 (simplify some `proc_macro` things) - #157883 (Remove strict invariant node_type on hir_type during ty privacy visit) - #157921 (trait solver: Resolve region vars in max universe) - #157960 (delegation: add support for infers in generics) - #157983 (Lift the same-signature restriction for `extern "tail"`) - #158053 (Optimize network address parser) - #158105 (Extract all instance shim variants into new `ShimKind` enum) - #158207 (Resolver: local/external split of `resolve_ident_in_module_non_globs_unadjusted` ) - #158279 (Follow goto and drop when linting unreachable code) - #157527 (Move derive tests into their dedicated folder) - #157807 (don't ice on non-lifetime binders under `-Zassumptions-on-binders`) - #158020 (Update mingw-w64 C toolchain) - #158039 (c-variadic: test that we use equality up to free lifetimes) - #158222 (format: ignore println newline in foreign format hints) - #158223 (Move target checking for #[lang] to the attribute parser) - #158252 (Use `cfg_select` in `std::os`) - #158257 ( fix escaping placeholder check in next solver normalization folder) - #158263 (Only load the feature list once in the entire resolver) - #158267 (FromUtf8Error::into_utf8_lossy better example and suggest use) - #158274 (triagebot: Stop pinging myself) - #158282 (slice_split_once: bounds check optimization note) - #158300 (Improve unknown crate_type diagnostic suggestions) - #158304 (mailmap: update mu001999) - #158309 (Update `rustc-literal-escaper` version to `0.0.8`) Failed merges: - #158256 (Avoid parser panics bubbling out to proc macros)

…ders-placeholder-ice, r=BoxyUwU trait solver: Resolve region vars in max universe Fixes rust-lang#157862 The ICE comes from computing the max universe of a region constraint after a region var has already been unified with something else. In that state, asking universe_of_lt directly can hit None and panic. This makes the region visitor opportunistically resolve ReVars first, then inspect the resolved region. Type and const infer vars stay as-is because this path already bails on non-region inference before it needs more structure from them.

Rollup of 35 pull requests Successful merges: - #158315 (`rust-analyzer` subtree update) - #158336 (Stop excluding `stdarch` test crates from `rust-src`) - #155739 (Add temporary scope to assert_eq and assert_ne) - #156885 (Fix misattributed type inference error span for index expressions) - #157271 (simplify some `proc_macro` things) - #157419 (move rustc_type_ir Term things to term_kind.rs) - #157883 (Remove strict invariant node_type on hir_type during ty privacy visit) - #157921 (trait solver: Resolve region vars in max universe) - #157960 (delegation: add support for infers in generics) - #157983 (Lift the same-signature restriction for `extern "tail"`) - #158053 (Optimize network address parser) - #158105 (Extract all instance shim variants into new `ShimKind` enum) - #158207 (Resolver: local/external split of `resolve_ident_in_module_non_globs_unadjusted` ) - #158279 (Follow goto and drop when linting unreachable code) - #157527 (Move derive tests into their dedicated folder) - #157807 (don't ice on non-lifetime binders under `-Zassumptions-on-binders`) - #157939 (Reorganize `tests/ui/issues` [8/N]) - #157946 (Make `char::is_private_use` and `char::is_assigned` unstably public) - #158003 (Reorganize `tests/ui/issues` [9/N]) - #158020 (Update mingw-w64 C toolchain) - #158039 (c-variadic: test that we use equality up to free lifetimes) - #158060 (Reorganize `tests/ui/issues` [10/N]) - #158222 (format: ignore println newline in foreign format hints) - #158223 (Move target checking for #[lang] to the attribute parser) - #158252 (Use `cfg_select` in `std::os`) - #158263 (Only load the feature list once in the entire resolver) - #158267 (FromUtf8Error::into_utf8_lossy better example and suggest use) - #158272 (Reorganize `tests/ui/issues` [13/N]) - #158274 (triagebot: Stop pinging myself) - #158282 (slice_split_once: bounds check optimization note) - #158300 (Improve unknown crate_type diagnostic suggestions) - #158304 (mailmap: update mu001999) - #158309 (Update `rustc-literal-escaper` version to `0.0.8`) - #158314 (Fix incorrect unsafe debug assertion in unchecked_div_exact) - #158326 (Add `io::ErrorKind::TooManyOpenFiles`)

…ders-placeholder-ice, r=BoxyUwU trait solver: Resolve region vars in max universe Fixes rust-lang#157862 The ICE comes from computing the max universe of a region constraint after a region var has already been unified with something else. In that state, asking universe_of_lt directly can hit None and panic. This makes the region visitor opportunistically resolve ReVars first, then inspect the resolved region. Type and const infer vars stay as-is because this path already bails on non-region inference before it needs more structure from them.

Rollup of 35 pull requests Successful merges: - #158315 (`rust-analyzer` subtree update) - #158336 (Stop excluding `stdarch` test crates from `rust-src`) - #155739 (Add temporary scope to assert_eq and assert_ne) - #156885 (Fix misattributed type inference error span for index expressions) - #157271 (simplify some `proc_macro` things) - #157419 (move rustc_type_ir Term things to term_kind.rs) - #157883 (Remove strict invariant node_type on hir_type during ty privacy visit) - #157921 (trait solver: Resolve region vars in max universe) - #157960 (delegation: add support for infers in generics) - #157983 (Lift the same-signature restriction for `extern "tail"`) - #158053 (Optimize network address parser) - #158105 (Extract all instance shim variants into new `ShimKind` enum) - #158207 (Resolver: local/external split of `resolve_ident_in_module_non_globs_unadjusted` ) - #158279 (Follow goto and drop when linting unreachable code) - #157527 (Move derive tests into their dedicated folder) - #157807 (don't ice on non-lifetime binders under `-Zassumptions-on-binders`) - #157939 (Reorganize `tests/ui/issues` [8/N]) - #157946 (Make `char::is_private_use` and `char::is_assigned` unstably public) - #158003 (Reorganize `tests/ui/issues` [9/N]) - #158020 (Update mingw-w64 C toolchain) - #158039 (c-variadic: test that we use equality up to free lifetimes) - #158060 (Reorganize `tests/ui/issues` [10/N]) - #158222 (format: ignore println newline in foreign format hints) - #158223 (Move target checking for #[lang] to the attribute parser) - #158252 (Use `cfg_select` in `std::os`) - #158263 (Only load the feature list once in the entire resolver) - #158267 (FromUtf8Error::into_utf8_lossy better example and suggest use) - #158272 (Reorganize `tests/ui/issues` [13/N]) - #158274 (triagebot: Stop pinging myself) - #158282 (slice_split_once: bounds check optimization note) - #158300 (Improve unknown crate_type diagnostic suggestions) - #158304 (mailmap: update mu001999) - #158309 (Update `rustc-literal-escaper` version to `0.0.8`) - #158314 (Fix incorrect unsafe debug assertion in unchecked_div_exact) - #158326 (Add `io::ErrorKind::TooManyOpenFiles`)

Rollup merge of #157921 - Dnreikronos:fix-assumptions-on-binders-placeholder-ice, r=BoxyUwU trait solver: Resolve region vars in max universe Fixes #157862 The ICE comes from computing the max universe of a region constraint after a region var has already been unified with something else. In that state, asking universe_of_lt directly can hit None and panic. This makes the region visitor opportunistically resolve ReVars first, then inspect the resolved region. Type and const infer vars stay as-is because this path already bails on non-region inference before it needs more structure from them.

JonathanBrouwer · 2026-06-25T10:54:01Z

@rust-timer build bc6d88d
Pef run for #158342

rust-timer · 2026-06-25T15:05:02Z

Finished benchmarking commit (bc6d88d): comparison URL.

Overall result: ✅ improvements - no action needed

Benchmarking means the PR may be perf-sensitive. Consider adding rollup=never if this change is not fit for rolling up.

@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.2%	[-0.2%, -0.2%]	1
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (secondary 3.5%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	3.5%	[3.5%, 3.5%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Cycles

Results (secondary 2.2%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	2.2%	[2.2%, 2.2%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Binary size

This perf run didn't have relevant results for this metric.

Bootstrap: 503.101s -> 505.578s (0.49%)
Artifact size: 353.51 MiB -> 353.60 MiB (0.02%)

Rollup of 35 pull requests Successful merges: - rust-lang/rust#158315 (`rust-analyzer` subtree update) - rust-lang/rust#158336 (Stop excluding `stdarch` test crates from `rust-src`) - rust-lang/rust#155739 (Add temporary scope to assert_eq and assert_ne) - rust-lang/rust#156885 (Fix misattributed type inference error span for index expressions) - rust-lang/rust#157271 (simplify some `proc_macro` things) - rust-lang/rust#157419 (move rustc_type_ir Term things to term_kind.rs) - rust-lang/rust#157883 (Remove strict invariant node_type on hir_type during ty privacy visit) - rust-lang/rust#157921 (trait solver: Resolve region vars in max universe) - rust-lang/rust#157960 (delegation: add support for infers in generics) - rust-lang/rust#157983 (Lift the same-signature restriction for `extern "tail"`) - rust-lang/rust#158053 (Optimize network address parser) - rust-lang/rust#158105 (Extract all instance shim variants into new `ShimKind` enum) - rust-lang/rust#158207 (Resolver: local/external split of `resolve_ident_in_module_non_globs_unadjusted` ) - rust-lang/rust#158279 (Follow goto and drop when linting unreachable code) - rust-lang/rust#157527 (Move derive tests into their dedicated folder) - rust-lang/rust#157807 (don't ice on non-lifetime binders under `-Zassumptions-on-binders`) - rust-lang/rust#157939 (Reorganize `tests/ui/issues` [8/N]) - rust-lang/rust#157946 (Make `char::is_private_use` and `char::is_assigned` unstably public) - rust-lang/rust#158003 (Reorganize `tests/ui/issues` [9/N]) - rust-lang/rust#158020 (Update mingw-w64 C toolchain) - rust-lang/rust#158039 (c-variadic: test that we use equality up to free lifetimes) - rust-lang/rust#158060 (Reorganize `tests/ui/issues` [10/N]) - rust-lang/rust#158222 (format: ignore println newline in foreign format hints) - rust-lang/rust#158223 (Move target checking for #[lang] to the attribute parser) - rust-lang/rust#158252 (Use `cfg_select` in `std::os`) - rust-lang/rust#158263 (Only load the feature list once in the entire resolver) - rust-lang/rust#158267 (FromUtf8Error::into_utf8_lossy better example and suggest use) - rust-lang/rust#158272 (Reorganize `tests/ui/issues` [13/N]) - rust-lang/rust#158274 (triagebot: Stop pinging myself) - rust-lang/rust#158282 (slice_split_once: bounds check optimization note) - rust-lang/rust#158300 (Improve unknown crate_type diagnostic suggestions) - rust-lang/rust#158304 (mailmap: update mu001999) - rust-lang/rust#158309 (Update `rustc-literal-escaper` version to `0.0.8`) - rust-lang/rust#158314 (Fix incorrect unsafe debug assertion in unchecked_div_exact) - rust-lang/rust#158326 (Add `io::ErrorKind::TooManyOpenFiles`)

rustbot assigned wesleywiser Jun 15, 2026

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 Jun 15, 2026

rustbot assigned BoxyUwU and unassigned wesleywiser Jun 15, 2026

rustbot added the WG-trait-system-refactor The Rustc Trait System Refactor Initiative (-Znext-solver) label Jun 16, 2026

Dnreikronos changed the title ~~fix ice from unwrapping universe_of_lt on an already-resolved region var~~ trait solver: Capture binder region constraints while relating Jun 16, 2026

Dnreikronos force-pushed the fix-assumptions-on-binders-placeholder-ice branch from 2dc4357 to a390036 Compare June 16, 2026 18:42

Dnreikronos changed the title ~~trait solver: Capture binder region constraints while relating~~ trait solver: Resolve region vars in max universe Jun 16, 2026

BoxyUwU reviewed Jun 19, 2026

View reviewed changes

Resolve vars before computing max universe

3bb0a1c

This comment has been minimized.

Sign in to view

Format max universe visitor

a44ab8d

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-review Status: Awaiting review from the assignee but also interested parties. labels Jun 23, 2026

JonathanBrouwer mentioned this pull request Jun 23, 2026

Rollup of 16 pull requests #158301

Closed

JonathanBrouwer mentioned this pull request Jun 23, 2026

Rollup of 23 pull requests #158316

Closed

JonathanBrouwer mentioned this pull request Jun 23, 2026

Rollup of 23 pull requests #158324

Closed

jhpratt mentioned this pull request Jun 24, 2026

Rollup of 27 pull requests #158335

Closed

This was referenced Jun 24, 2026

Rollup of 34 pull requests #158337

Closed

Rollup of 35 pull requests #158342

Merged

JonathanBrouwer mentioned this pull request Jun 24, 2026

[CI] Rollup of 35 pull requests #158354

Draft

rust-bors Bot merged commit 0b4e740 into rust-lang:main Jun 25, 2026
13 checks passed

rustbot added this to the 1.98.0 milestone Jun 25, 2026

This comment has been minimized.

Sign in to view

Uh oh!

Uh oh!

Conversation

Dnreikronos commented Jun 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Jun 15, 2026

Uh oh!

BoxyUwU commented Jun 15, 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!

BoxyUwU Jun 19, 2026

Choose a reason for hiding this comment

Uh oh!

BoxyUwU Jun 19, 2026

Choose a reason for hiding this comment

Uh oh!

Dnreikronos commented Jun 20, 2026

Uh oh!

This comment has been minimized.

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

Uh oh!

Uh oh!

BoxyUwU commented Jun 23, 2026

Uh oh!

rust-bors Bot commented Jun 23, 2026

Uh oh!

Uh oh!

JonathanBrouwer commented Jun 25, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Jun 25, 2026

Overall result: ✅ improvements - no action needed

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

Dnreikronos commented Jun 15, 2026 •

edited

Loading

BoxyUwU left a comment •

edited by rustbot

Loading

Dnreikronos commented Jun 20, 2026 •

edited

Loading