Expand weak alias types before collecting constrained/referenced late bound regions + refactorings#121344

fmease · 2024-02-20T14:12:28Z

Fixes #114220.
Follow-up to #120780.

fmease · 2024-02-20T14:15:11Z

compiler/rustc_middle/src/query/mod.rs

-    query normalize_weak_ty(
-        goal: CanonicalProjectionGoal<'tcx>
+    /// <div class="warning">Do not call this query directly: Invoke `normalize` instead.</div>
+    query normalize_canonicalized_weak_ty(


As discussed I've renamed those queries to something “less appealing” (well I could also add a leading _ for good measure :P) but since I actually chose to use the term expand for the new normalization routine, I can revert this change since their names no longer really clash.

fmease · 2024-02-20T14:16:56Z

compiler/rustc_middle/src/ty/util.rs

+    ///
+    /// [weak]: ty::Weak
+    pub fn expand_weak_alias_tys<T: TypeFoldable<TyCtxt<'tcx>>>(self, value: T) -> T {
+        value.fold_with(&mut WeakAliasTypeExpander { tcx: self, depth: 0 })


I wonder if I need to track binders. I mean I didn't do that before either when I did the equivalent of expand_weak_alias_tys inline in an ad hoc fashion but yeah, feels weird to not track binders even though instantiate should take care of that (it does at least shift_bound_vars).

Binders do not matter here. You always instantiate the args from the same level, and you do not invoke the full projection normalizer at all.

fmease · 2024-02-20T14:18:51Z

compiler/rustc_middle/src/ty/util.rs


+struct WeakAliasTypeExpander<'tcx> {
+    tcx: TyCtxt<'tcx>,
+    depth: usize,


I'd like to note that I'm still using a plain depth counter over a hash set of DefIds to detect overflows since that feels more memory efficient to me.

fmease · 2024-02-20T14:20:55Z

compiler/rustc_middle/src/ty/util.rs

+            return ty.super_fold_with(self);
+        };
+        if !self.tcx.recursion_limit().value_within_limit(self.depth) {
+            let guar = self.tcx.dcx().delayed_bug("overflow expanding weak alias type");


I'm still delaying a bug instead of reporting an overflow error since we don't have access to a Span and neither do any the callers of expand_weak_alias_tys and it would be a bit foolish to use a dummy span for this since it wouldn't get deduplicated with the other overflow errors and would actually show up in the user's stderr.

compiler/rustc_trait_selection/src/traits/normalize.rs

fmease · 2024-02-20T14:44:20Z

compiler/rustc_middle/src/ty/util.rs

+    /// Peel off all [weak alias types] in this type until there are none left.
+    ///
+    /// This only expands weak alias types in “head” / outermost positions. It can
+    /// be used over [expand_weak_alias_tys] as an optimization in situations where


“optimization modulo overflow behavior” I guess

oli-obk · 2024-02-20T15:00:01Z

compiler/rustc_middle/src/ty/visit.rs

 }

 impl LateBoundRegionsCollector {
-    fn new(tcx: TyCtxt<'tcx>, just_constrained: bool) -> Self {


huh why did this not cause a lint?

Ah because it turns out my commits weren't as cleanly separated as I thought they would be ^^'. An earlier commit adds this unused binding and a later one removes it again. Let me fix that real quick.

oli-obk · 2024-02-20T15:17:44Z

Yea, I like the way this looks

oli-obk · 2024-02-20T15:18:06Z

@bors r+

bors · 2024-02-20T15:18:08Z

📌 Commit 9bf9861 has been approved by oli-obk

It is now in the queue for this repository.

fmease · 2024-02-20T16:27:42Z

Fixing up the commit history
@bors r-

…late bound regions

… in impls

fmease · 2024-02-20T16:32:47Z

@bors r=oli-obk

bors · 2024-02-20T16:32:49Z

📌 Commit f515f99 has been approved by oli-obk

It is now in the queue for this repository.

…iaskrgr Rollup of 7 pull requests Successful merges: - rust-lang#119203 (Correct the simd_masked_{load,store} intrinsic docs) - rust-lang#121277 (Refactor trait implementations in `core::convert::num`.) - rust-lang#121322 (Don't ICE when hitting overflow limit in fulfillment loop in next solver) - rust-lang#121323 (Don't use raw parameter types in `find_builder_fn`) - rust-lang#121344 (Expand weak alias types before collecting constrained/referenced late bound regions + refactorings) - rust-lang#121350 (Fix stray trait mismatch in `resolve_associated_item` for `AsyncFn`) - rust-lang#121352 (docs: add missing "the" to `str::strip_prefix` doc) Failed merges: - rust-lang#121340 (bootstrap: apply most of clippy's suggestions) r? `@ghost` `@rustbot` modify labels: rollup

Rollup merge of rust-lang#121344 - fmease:lta-constr-by-input, r=oli-obk Expand weak alias types before collecting constrained/referenced late bound regions + refactorings Fixes rust-lang#114220. Follow-up to rust-lang#120780. r? `@oli-obk`

…cts, r=compiler-errors Also expand weak alias tys inside consts inside `expand_weak_alias_tys` Ever since rust-lang#121344 has been merged, I couldn't let go of the fear that I might've slipped a tiny bug into rustc (:P). Checking the type flags of the `Const` is strictly more correct than only checking the ones of the `Const`'s `Ty`. I don't think it's possible to trigger an ICE rn (i.e., one of the two `bug!("unexpected weak alias type")` I added in branches where `expand_weak_alias_tys` should've expanded *all* weak alias tys) because presently const exprs aren't allowed to capture late-bound vars. To be future-proof however, we should iron this out. A possible reproducer would be the following if I'm not mistaken (currently fails to compile due to the aforementioned restriction): ```rs #![feature(lazy_type_alias, adt_const_params, generic_const_exprs)] type F = for<'a> fn(A<{ S::<Weak<'a>>(loop {}) }>) -> &'a (); type A<const N: S<Weak<'static>>> = (); #[derive(PartialEq, Eq, std::marker::ConstParamTy)] struct S<T>(T); type Weak<'a> = &'a (); ``` Whether a late-bound region should actually be considered constrained by a const expr is a separate question — one which we don't need to answer until / unless we actually allow them in such contexts (probable answer: only inside the return exprs of a block but not inside the stmts). r? oli-obk (he's not available rn but that's fine) or types or compiler

Rollup merge of rust-lang#124990 - fmease:expand-weak-aliases-within-cts, r=compiler-errors Also expand weak alias tys inside consts inside `expand_weak_alias_tys` Ever since rust-lang#121344 has been merged, I couldn't let go of the fear that I might've slipped a tiny bug into rustc (:P). Checking the type flags of the `Const` is strictly more correct than only checking the ones of the `Const`'s `Ty`. I don't think it's possible to trigger an ICE rn (i.e., one of the two `bug!("unexpected weak alias type")` I added in branches where `expand_weak_alias_tys` should've expanded *all* weak alias tys) because presently const exprs aren't allowed to capture late-bound vars. To be future-proof however, we should iron this out. A possible reproducer would be the following if I'm not mistaken (currently fails to compile due to the aforementioned restriction): ```rs #![feature(lazy_type_alias, adt_const_params, generic_const_exprs)] type F = for<'a> fn(A<{ S::<Weak<'a>>(loop {}) }>) -> &'a (); type A<const N: S<Weak<'static>>> = (); #[derive(PartialEq, Eq, std::marker::ConstParamTy)] struct S<T>(T); type Weak<'a> = &'a (); ``` Whether a late-bound region should actually be considered constrained by a const expr is a separate question — one which we don't need to answer until / unless we actually allow them in such contexts (probable answer: only inside the return exprs of a block but not inside the stmts). r? oli-obk (he's not available rn but that's fine) or types or compiler

fmease added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-types Relevant to the types team, which will review and decide on the PR/issue. F-lazy_type_alias `#![feature(lazy_type_alias)]` labels Feb 20, 2024

rustbot assigned oli-obk Feb 20, 2024

fmease commented Feb 20, 2024

View reviewed changes

compiler/rustc_trait_selection/src/traits/normalize.rs Show resolved Hide resolved

fmease changed the title ~~Expand weak alias types before collecting constrained and referenced late bound regions + refactorings~~ Expand weak alias types before collecting constrained & referenced late bound regions + refactorings Feb 20, 2024

fmease changed the title ~~Expand weak alias types before collecting constrained & referenced late bound regions + refactorings~~ Expand weak alias types before collecting constrained/referenced late bound regions + refactorings Feb 20, 2024

fmease commented Feb 20, 2024

View reviewed changes

oli-obk reviewed Feb 20, 2024

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 Feb 20, 2024

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 Feb 20, 2024

fmease added 5 commits February 20, 2024 17:30

Rename some normalization-related items

05ce209

Introduce expand_weak_alias_tys

515d805

Expand weak alias types before collecting constrained and referenced …

da01cce

…late bound regions

Use expand_weak_alias_tys when collecting constrained generics params…

1b3df6f

… in impls

Move the peeling function for weak alias types

f515f99

fmease force-pushed the lta-constr-by-input branch from 9bf9861 to f515f99 Compare February 20, 2024 16:32

bors removed the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Feb 20, 2024

bors added the S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. label Feb 20, 2024

matthiaskrgr mentioned this pull request Feb 20, 2024

Rollup of 7 pull requests #121356

Merged

bors merged commit 532b3ea into rust-lang:master Feb 21, 2024

rustbot added this to the 1.78.0 milestone Feb 21, 2024

fmease deleted the lta-constr-by-input branch February 21, 2024 04:08

fmease mentioned this pull request Feb 21, 2024

Tracking issue for lazy type aliases #112792

Open

6 tasks

fmease mentioned this pull request Mar 11, 2024

Treat weak alias types more like ADTs when computing implied bounds #122340

Closed

fmease mentioned this pull request May 10, 2024

Also expand weak alias tys inside consts inside expand_weak_alias_tys #124990

Merged

fmease mentioned this pull request Apr 24, 2025

Free type aliases should not require all generic parameters to be used #140230

Open

moxian mentioned this pull request Jun 15, 2025

ICE: overflow expanding free alias type #142419

Closed

fmease added the T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. label Oct 9, 2025

Uh oh!

Conversation

fmease commented Feb 20, 2024

Uh oh!

fmease Feb 20, 2024

Choose a reason for hiding this comment

Uh oh!

fmease Feb 20, 2024 • 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!

oli-obk Feb 20, 2024

Choose a reason for hiding this comment

Uh oh!

fmease Feb 20, 2024

Choose a reason for hiding this comment

Uh oh!

fmease Feb 20, 2024 • 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!

Uh oh!

fmease Feb 20, 2024

Choose a reason for hiding this comment

Uh oh!

oli-obk Feb 20, 2024

Choose a reason for hiding this comment

Uh oh!

fmease Feb 20, 2024 • 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!

oli-obk commented Feb 20, 2024

Uh oh!

oli-obk commented Feb 20, 2024

Uh oh!

bors commented Feb 20, 2024

Uh oh!

fmease commented Feb 20, 2024

Uh oh!

fmease commented Feb 20, 2024

Uh oh!

bors commented Feb 20, 2024

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

fmease Feb 20, 2024 •

edited

Loading

fmease Feb 20, 2024 •

edited

Loading

fmease Feb 20, 2024 •

edited

Loading