traits: Represent live alias arguments as bitsets by Dnreikronos · Pull Request #160936 · rust-lang/rust

Dnreikronos · 2026-08-11T18:21:55Z

Split out of #160212 per review / Zulip: https://rust-lang.zulipchat.com/#narrow/channel/364551-t-types.2Ftrait-system-refactor/topic/rigid.20aliases.20in.20region.20handling/near/615557841

live_args_for_alias_from_outlives_bounds and args_known_to_outlive_alias_params now return identity arg indices (DenseBitSet) instead of EarlyBinder<GenericArg>. Callers just do args[idx].visit_with(...), so we stop laundering rigidness through binder instantiate. Also drops the old BitSet FIXME.

imo this is worth doing on its own even without the ICE fix. every time we shoved identity params through EarlyBinder we were writing down something we don't actually know, and this module is only going to grow. better to make the invariant explicit now than keep paying for it later.

No behavioral change intended. the rigid-alias ICE / extract_verify_if_eq bits stay on #160212. btw if that one should stack on this instead of staying independent, lmk and I'll rebase it asap.

Store identity argument indices instead of bound generic arguments so callers can index concrete alias arguments without changing rigidness through instantiation.

rustbot · 2026-08-11T18:21:58Z

These commits modify the Cargo.lock file. Unintentional changes to Cargo.lock can be introduced when switching branches and rebasing PRs.

If this was unintentional then you should revert the changes before this PR is merged.
Otherwise, you can ignore this comment.

rustbot · 2026-08-11T18:22:01Z

r? @JohnTitor

rustbot has assigned @JohnTitor.
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 18 candidates

Dnreikronos · 2026-08-11T18:29:14Z

@lcnr hey, this is the bitset-only split from #160212 / the Zulip thread: https://rust-lang.zulipchat.com/#narrow/channel/364551-t-types.2Ftrait-system-refactor/topic/rigid.20aliases.20in.20region.20handling/near/615557841

fyi no matcher / ICE changes here. just arg indices via DenseBitSet, so callers do args[idx] instead of EarlyBinder + instantiate. also kills the BitSet FIXME.

imo this is the boring half and should be easy to land first. the interesting open question (proper matching vs underapproximating in extract_verify_if_eq) stays on #160212. lmk if anything looks off or if you want me to rebase that one on top of this asap.

panstromek · 2026-08-11T19:05:28Z

@bors try @rust-timer queue

I assume we want to test perf on this, this seems like it could have perf implications.

traits: Represent live alias arguments as bitsets

rust-bors · 2026-08-11T20:05:07Z

☀️ Try build successful (CI)
Build commit: eb7fea4 (eb7fea435c224921ea27002d0fa50f308f2a2e00)
Base parent: e64c8a6 (e64c8a664d9da54fc239cd4404cbf67f0d624326)

rust-timer · 2026-08-11T20:47:34Z

Finished benchmarking commit (eb7fea4): comparison URL.

Overall result: no relevant changes - 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

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

Max RSS (memory usage)

Results (secondary 2.7%)

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)	5.2%	[0.5%, 12.1%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.5%	[-0.7%, -0.4%]	3
All ❌✅ (primary)	-	-	0

Cycles

Results (secondary -0.2%)

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)	0.9%	[0.5%, 2.3%]	8
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.9%	[-1.9%, -0.4%]	13
All ❌✅ (primary)	-	-	0

Binary size

Results (secondary -0.1%)

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)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.1%	[-0.2%, -0.0%]	16
All ❌✅ (primary)	-	-	0

Bootstrap: 455.693s -> 454.48s (-0.27%)
Artifact size: 396.39 MiB -> 396.46 MiB (0.02%)

adwinwhite · 2026-08-12T07:30:00Z

Would it be better to use an index newtype since it's on public API? Unsure it matters much since it's actually only used in the same module.

lcnr · 2026-08-12T10:00:11Z

Would it be better to use an index newtype since it's on public API? Unsure it matters much since it's actually only used in the same module.

either seems fine to me personally 🤷

Match the existing params_in_repr / unsizing_params convention; the compiler already treats generic arg counts as u32-sized.

jackh726 · 2026-08-16T01:24:55Z

        };
    }
-    live_args.map(|c| ty::EarlyBinder::bind(tcx, c.into_iter().collect()))
+    live_args


Honestly, this can just be DenseBitSet::new_filled at this point (and similarly for the None cases of live_args_for_outlives_clause)...then the restrict loops start with a filled set and just take the intersection each time.

(No outlives bounds means we consider everything as potentially live.)

Though: we may need to filter the bivariant args in the loop for the Some case though.

View changes since the review

Yeah, makes sense. I just changed this. When there are no outlives bounds, or a caller clause does not apply, that source does not narrow down the live args. So I start with all args included and let each real restriction intersect the set. That is why DenseBitSet::new_filled fits better than using None to mean "no restriction".

I also changed the query and metadata entry to return the bitset directly. I kept the bivariant check when walking the final set, since a filled set includes those args but we still should not visit them. That was the edge case you pointed out.

I think this is easier to follow than keeping the Option branches around.

Treat missing outlives information as no restriction so all sources can be intersected uniformly. Keep bivariant alias arguments out of the final region walk.

adwinwhite · 2026-09-04T09:06:46Z

I thought this was already merged 🤣

r? me

@bors r=lcnr,jackh726,adwinwhite rollup

rust-bors · 2026-09-04T09:06:49Z

📌 Commit c4f07b8 has been approved by lcnr,jackh726,adwinwhite

It is now in the queue for this repository.

…t, r=lcnr,jackh726,adwinwhite traits: Represent live alias arguments as bitsets Split out of rust-lang#160212 per review / Zulip: https://rust-lang.zulipchat.com/#narrow/channel/364551-t-types.2Ftrait-system-refactor/topic/rigid.20aliases.20in.20region.20handling/near/615557841 `live_args_for_alias_from_outlives_bounds` and `args_known_to_outlive_alias_params` now return identity arg indices (`DenseBitSet`) instead of `EarlyBinder<GenericArg>`. Callers just do `args[idx].visit_with(...)`, so we stop laundering rigidness through binder instantiate. Also drops the old BitSet FIXME. imo this is worth doing on its own even without the ICE fix. every time we shoved identity params through `EarlyBinder` we were writing down something we don't actually know, and this module is only going to grow. better to make the invariant explicit now than keep paying for it later. No behavioral change intended. the rigid-alias ICE / `extract_verify_if_eq` bits stay on rust-lang#160212. btw if that one should stack on this instead of staying independent, lmk and I'll rebase it asap.

Rollup of 25 pull requests Successful merges: - #159074 ([PAC] FnAbi, llvm.ptrauth.resign and Session API change (2/8)) - #159792 (A more readable debug map for IndexMaps) - #161895 (std::sys::pal::sgx: fix mismatched alloc/free alignment) - #161900 (bootstrap: Include feature-gated items in bootstrap tool docs) - #161940 (Promote `wasm32-wasip3` to a tier 2 target) - #162072 (Add new Tier-3 target: `powerpc64-sony-ps3`) - #162179 (type system const items via direct rhs) - #162277 (Introduce `rustc_middle::middel::resolve`) - #162285 (box: fixup map/try_map deallocate calls) - #162286 (string: don't unwind prematurely) - #162289 (alloc: a bunch of safety comments) - #162292 (Update `askama` version to `0.16.1`) - #160509 (Remove `RegionExt`; move methods to `Region` in `rustc_type_ir`) - #160906 (Suggest usize instead of placeholder type for array length constants) - #160936 (traits: Represent live alias arguments as bitsets) - #161400 (Improve diagnostics for references to closures) - #161656 (Suggest mutable references for FnMut closure arguments) - #161711 (Add more splat fn type tests) - #161786 (Make `tcx.def_id_partial_cmp` public) - #161953 (sanitizers: Implicitly disable mutually exclusive sanitizers) - #162155 (add suggestion for `rustc_allowed_through_unstable_modules` attribute) - #162212 (Implement `Rng` for `Box`) - #162246 (Fix incorrect meta span) - #162266 (std: fix typo) - #162291 (Add regression test from 1.98.1)

…t, r=lcnr,jackh726,adwinwhite traits: Represent live alias arguments as bitsets Split out of rust-lang#160212 per review / Zulip: https://rust-lang.zulipchat.com/#narrow/channel/364551-t-types.2Ftrait-system-refactor/topic/rigid.20aliases.20in.20region.20handling/near/615557841 `live_args_for_alias_from_outlives_bounds` and `args_known_to_outlive_alias_params` now return identity arg indices (`DenseBitSet`) instead of `EarlyBinder<GenericArg>`. Callers just do `args[idx].visit_with(...)`, so we stop laundering rigidness through binder instantiate. Also drops the old BitSet FIXME. imo this is worth doing on its own even without the ICE fix. every time we shoved identity params through `EarlyBinder` we were writing down something we don't actually know, and this module is only going to grow. better to make the invariant explicit now than keep paying for it later. No behavioral change intended. the rigid-alias ICE / `extract_verify_if_eq` bits stay on rust-lang#160212. btw if that one should stack on this instead of staying independent, lmk and I'll rebase it asap.

Rollup of 25 pull requests Successful merges: - #159074 ([PAC] FnAbi, llvm.ptrauth.resign and Session API change (2/8)) - #159792 (A more readable debug map for IndexMaps) - #160745 (make closures act like MaybeDangling) - #161895 (std::sys::pal::sgx: fix mismatched alloc/free alignment) - #161940 (Promote `wasm32-wasip3` to a tier 2 target) - #162072 (Add new Tier-3 target: `powerpc64-sony-ps3`) - #162179 (type system const items via direct rhs) - #162277 (Introduce `rustc_middle::middel::resolve`) - #162285 (box: fixup map/try_map deallocate calls) - #162286 (string: don't unwind prematurely) - #162289 (alloc: a bunch of safety comments) - #162292 (Update `askama` version to `0.16.1`) - #160509 (Remove `RegionExt`; move methods to `Region` in `rustc_type_ir`) - #160906 (Suggest usize instead of placeholder type for array length constants) - #160936 (traits: Represent live alias arguments as bitsets) - #161400 (Improve diagnostics for references to closures) - #161656 (Suggest mutable references for FnMut closure arguments) - #161711 (Add more splat fn type tests) - #161786 (Make `tcx.def_id_partial_cmp` public) - #161953 (sanitizers: Implicitly disable mutually exclusive sanitizers) - #162155 (add suggestion for `rustc_allowed_through_unstable_modules` attribute) - #162212 (Implement `Rng` for `Box`) - #162246 (Fix incorrect meta span) - #162266 (std: fix typo) - #162291 (Add regression test from 1.98.1)

…t, r=lcnr,jackh726,adwinwhite traits: Represent live alias arguments as bitsets Split out of rust-lang#160212 per review / Zulip: https://rust-lang.zulipchat.com/#narrow/channel/364551-t-types.2Ftrait-system-refactor/topic/rigid.20aliases.20in.20region.20handling/near/615557841 `live_args_for_alias_from_outlives_bounds` and `args_known_to_outlive_alias_params` now return identity arg indices (`DenseBitSet`) instead of `EarlyBinder<GenericArg>`. Callers just do `args[idx].visit_with(...)`, so we stop laundering rigidness through binder instantiate. Also drops the old BitSet FIXME. imo this is worth doing on its own even without the ICE fix. every time we shoved identity params through `EarlyBinder` we were writing down something we don't actually know, and this module is only going to grow. better to make the invariant explicit now than keep paying for it later. No behavioral change intended. the rigid-alias ICE / `extract_verify_if_eq` bits stay on rust-lang#160212. btw if that one should stack on this instead of staying independent, lmk and I'll rebase it asap.

Rollup of 27 pull requests Successful merges: - #159074 ([PAC] FnAbi, llvm.ptrauth.resign and Session API change (2/8)) - #159792 (A more readable debug map for IndexMaps) - #160745 (make closures act like MaybeDangling) - #161940 (Promote `wasm32-wasip3` to a tier 2 target) - #162030 (Prevent `--test` to be used in `rustdoc-html` testsuite) - #162072 (Add new Tier-3 target: `powerpc64-sony-ps3`) - #162179 (type system const items via direct rhs) - #162262 (Avoid manually instantiating some binders in error reporting with `-Znext-solver`) - #162277 (Introduce `rustc_middle::middel::resolve`) - #162285 (box: fixup map/try_map deallocate calls) - #162286 (string: don't unwind prematurely) - #162289 (alloc: a bunch of safety comments) - #162290 (abby test DSL: AliasTyOutlivesViaEnv) - #162292 (Update `askama` version to `0.16.1`) - #160509 (Remove `RegionExt`; move methods to `Region` in `rustc_type_ir`) - #160906 (Suggest usize instead of placeholder type for array length constants) - #160936 (traits: Represent live alias arguments as bitsets) - #161400 (Improve diagnostics for references to closures) - #161656 (Suggest mutable references for FnMut closure arguments) - #161711 (Add more splat fn type tests) - #161786 (Make `tcx.def_id_partial_cmp` public) - #161953 (sanitizers: Implicitly disable mutually exclusive sanitizers) - #162155 (add suggestion for `rustc_allowed_through_unstable_modules` attribute) - #162212 (Implement `Rng` for `Box`) - #162246 (Fix incorrect meta span) - #162266 (std: fix typo) - #162291 (Add regression test from 1.98.1)

…t, r=lcnr,jackh726,adwinwhite traits: Represent live alias arguments as bitsets Split out of rust-lang#160212 per review / Zulip: https://rust-lang.zulipchat.com/#narrow/channel/364551-t-types.2Ftrait-system-refactor/topic/rigid.20aliases.20in.20region.20handling/near/615557841 `live_args_for_alias_from_outlives_bounds` and `args_known_to_outlive_alias_params` now return identity arg indices (`DenseBitSet`) instead of `EarlyBinder<GenericArg>`. Callers just do `args[idx].visit_with(...)`, so we stop laundering rigidness through binder instantiate. Also drops the old BitSet FIXME. imo this is worth doing on its own even without the ICE fix. every time we shoved identity params through `EarlyBinder` we were writing down something we don't actually know, and this module is only going to grow. better to make the invariant explicit now than keep paying for it later. No behavioral change intended. the rigid-alias ICE / `extract_verify_if_eq` bits stay on rust-lang#160212. btw if that one should stack on this instead of staying independent, lmk and I'll rebase it asap.

…uwer Rollup of 25 pull requests Successful merges: - #162072 (Add new Tier-3 target: `powerpc64-sony-ps3`) - #162179 (type system const items via direct rhs) - #162255 (let people change rustc-dev-guide in-tree) - #162262 (Avoid manually instantiating some binders in error reporting with `-Znext-solver`) - #162277 (Introduce `rustc_middle::middel::resolve`) - #162285 (box: fixup map/try_map deallocate calls) - #162286 (string: don't unwind prematurely) - #162289 (alloc: a bunch of safety comments) - #162290 (abby test DSL: AliasTyOutlivesViaEnv) - #162292 (Update `askama` version to `0.16.1`) - #160509 (Remove `RegionExt`; move methods to `Region` in `rustc_type_ir`) - #160906 (Suggest usize instead of placeholder type for array length constants) - #160936 (traits: Represent live alias arguments as bitsets) - #161394 (fix `is_homogeneous_aggregate`: use unit size, not total size ) - #161400 (Improve diagnostics for references to closures) - #161656 (Suggest mutable references for FnMut closure arguments) - #161711 (Add more splat fn type tests) - #161786 (Make `tcx.def_id_partial_cmp` public) - #161953 (sanitizers: Implicitly disable mutually exclusive sanitizers) - #162155 (add suggestion for `rustc_allowed_through_unstable_modules` attribute) - #162212 (Implement `Rng` for `Box`) - #162246 (Fix incorrect meta span) - #162266 (std: fix typo) - #162291 (Add regression test from 1.98.1) - #162303 (Fix a minor spelling error in `library/std/src/sys/fs/unix.rs`)

…t, r=lcnr,jackh726,adwinwhite traits: Represent live alias arguments as bitsets Split out of rust-lang#160212 per review / Zulip: https://rust-lang.zulipchat.com/#narrow/channel/364551-t-types.2Ftrait-system-refactor/topic/rigid.20aliases.20in.20region.20handling/near/615557841 `live_args_for_alias_from_outlives_bounds` and `args_known_to_outlive_alias_params` now return identity arg indices (`DenseBitSet`) instead of `EarlyBinder<GenericArg>`. Callers just do `args[idx].visit_with(...)`, so we stop laundering rigidness through binder instantiate. Also drops the old BitSet FIXME. imo this is worth doing on its own even without the ICE fix. every time we shoved identity params through `EarlyBinder` we were writing down something we don't actually know, and this module is only going to grow. better to make the invariant explicit now than keep paying for it later. No behavioral change intended. the rigid-alias ICE / `extract_verify_if_eq` bits stay on rust-lang#160212. btw if that one should stack on this instead of staying independent, lmk and I'll rebase it asap.

…uwer Rollup of 25 pull requests Successful merges: - #162072 (Add new Tier-3 target: `powerpc64-sony-ps3`) - #162179 (type system const items via direct rhs) - #162255 (let people change rustc-dev-guide in-tree) - #162262 (Avoid manually instantiating some binders in error reporting with `-Znext-solver`) - #162277 (Introduce `rustc_middle::middel::resolve`) - #162285 (box: fixup map/try_map deallocate calls) - #162286 (string: don't unwind prematurely) - #162289 (alloc: a bunch of safety comments) - #162290 (abby test DSL: AliasTyOutlivesViaEnv) - #162292 (Update `askama` version to `0.16.1`) - #160509 (Remove `RegionExt`; move methods to `Region` in `rustc_type_ir`) - #160906 (Suggest usize instead of placeholder type for array length constants) - #160936 (traits: Represent live alias arguments as bitsets) - #161394 (fix `is_homogeneous_aggregate`: use unit size, not total size ) - #161400 (Improve diagnostics for references to closures) - #161656 (Suggest mutable references for FnMut closure arguments) - #161711 (Add more splat fn type tests) - #161786 (Make `tcx.def_id_partial_cmp` public) - #161953 (sanitizers: Implicitly disable mutually exclusive sanitizers) - #162155 (add suggestion for `rustc_allowed_through_unstable_modules` attribute) - #162212 (Implement `Rng` for `Box`) - #162246 (Fix incorrect meta span) - #162266 (std: fix typo) - #162291 (Add regression test from 1.98.1) - #162303 (Fix a minor spelling error in `library/std/src/sys/fs/unix.rs`)

Represent live alias arguments as bitsets

32aecb0

Store identity argument indices instead of bound generic arguments so callers can index concrete alias arguments without changing rigidness through instantiation.

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 Aug 11, 2026

rustbot assigned JohnTitor Aug 11, 2026

Dnreikronos mentioned this pull request Aug 11, 2026

traits: Fix rigid alias liveness matching #160212

Open

This comment has been minimized.

Sign in to view

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Aug 11, 2026

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Aug 11, 2026

Auto merge of #160936 - Dnreikronos:traits/live_args_bitset, r=<try>

eb7fea4

traits: Represent live alias arguments as bitsets

This comment has been minimized.

Sign in to view

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Aug 11, 2026

lcnr reviewed Aug 12, 2026

View reviewed changes

Comment thread compiler/rustc_metadata/src/rmeta/parameterized.rs Outdated

Use DenseBitSet<u32> for live alias args

9d3a26d

Match the existing params_in_repr / unsizing_params convention; the compiler already treats generic arg counts as u32-sized.

This comment was marked as resolved.

Sign in to view

oli-obk closed this Aug 13, 2026

rustbot removed the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Aug 13, 2026

oli-obk reopened this Aug 13, 2026

rustbot added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Aug 13, 2026

jackh726 reviewed Aug 16, 2026

View reviewed changes

Use filled bitsets for alias liveness restrictions

c4f07b8

Treat missing outlives information as no restriction so all sources can be intersected uniformly. Keep bivariant alias arguments out of the final region walk.

rustbot assigned adwinwhite and unassigned JohnTitor Sep 4, 2026

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 Sep 4, 2026

JonathanBrouwer mentioned this pull request Sep 4, 2026

Rollup of 13 pull requests #162293

Closed

GuillaumeGomez mentioned this pull request Sep 4, 2026

Rollup of 25 pull requests #162297

Closed

GuillaumeGomez mentioned this pull request Sep 4, 2026

Rollup of 25 pull requests #162298

Closed

GuillaumeGomez mentioned this pull request Sep 4, 2026

Rollup of 25 pull requests #162300

Closed

GuillaumeGomez mentioned this pull request Sep 4, 2026

Rollup of 28 pull requests #162307

Closed

GuillaumeGomez mentioned this pull request Sep 4, 2026

Rollup of 27 pull requests #162308

Closed

JonathanBrouwer mentioned this pull request Sep 4, 2026

Rollup of 25 pull requests #162310

Open

Zalathar mentioned this pull request Sep 5, 2026

Rollup of 20 pull requests #162317

Closed

Uh oh!

Uh oh!

Conversation

Dnreikronos commented Aug 11, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Aug 11, 2026

Uh oh!

rustbot commented Aug 11, 2026

Uh oh!

Dnreikronos commented Aug 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

panstromek commented Aug 11, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors Bot commented Aug 11, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Aug 11, 2026

Overall result: no relevant changes - no action needed

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Uh oh!

adwinwhite commented Aug 12, 2026

Uh oh!

Uh oh!

lcnr commented Aug 12, 2026

Uh oh!

This comment was marked as resolved.

jackh726 Aug 16, 2026 • 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 Aug 16, 2026

Choose a reason for hiding this comment

Uh oh!

adwinwhite commented Sep 4, 2026

Uh oh!

rust-bors Bot commented Sep 4, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

9 participants

Dnreikronos commented Aug 11, 2026 •

edited by rustbot

Loading

Dnreikronos commented Aug 11, 2026 •

edited

Loading

jackh726 Aug 16, 2026 •

edited by rustbot

Loading