Make alias-eq have a relation direction (and rename it to alias-relate) #109462

compiler-errors · 2023-03-21T22:45:53Z

Emitting an "alias-eq" is too strict in some situations, since we don't always want strict equality between a projection and rigid ty. Adds a relation direction.

I could probably just reuse this RelationDir -- happy to uplift that struct into middle and use that instead, but I didn't feel compelled to... 🤷
Some of the matching in compute_alias_relate_goal is a bit verbose -- I guess I could simplify it by using At::relate and mapping the relation-dir to a variance.
Alternatively, I coulld simplify things by making more helper functions on EvalCtxt (e.g. EvalCtxt::relate_with_direction(T, T) that also does the nested goal registration). No preference.

r? @lcnr cc @BoxyUwU though boxy can claim it if she wants
NOTE: first commit is all the changes, the second is just renaming stuff

rustbot · 2023-03-21T22:46:02Z

Some changes occurred in src/tools/clippy

cc @rust-lang/clippy

Some changes occurred to the core trait solver

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

lcnr

for caching it would be better if we only have AliasRelateDir::Eq and AliasRelateDir::Sub, or maybe in that case AliasRelateMode::Eq and AliasRelateMode::Sub, and instead of flipping the direction, flipping the order of the aliases 🤔

apart from that and nits 👍

compiler/rustc_infer/src/infer/nll_relate/mod.rs

compiler/rustc_trait_selection/src/solve/mod.rs

compiler/rustc_infer/src/infer/combine.rs

compiler/rustc_infer/src/infer/glb.rs

BoxyUwU · 2023-03-23T03:43:18Z

compiler/rustc_middle/src/ty/mod.rs

    ///
    /// FIXME: rename `AliasTy` to `AliasTerm` and make sure we correctly
    /// deal with constants.
-    pub fn to_alias_term_no_opaque(&self, tcx: TyCtxt<'tcx>) -> Option<AliasTy<'tcx>> {
+    pub fn to_projection_term(&self, tcx: TyCtxt<'tcx>) -> Option<AliasTy<'tcx>> {


I'm not a huge fan of this rename, when we start representing free type aliases in Ty I would expect this function to return Some for them. It's also not really accurate for ConstKind::Unevaluated which represents anon consts which are sort of like free consts (although nothing about normalization or equality of ConstKind::Unevaluated is really implemented right for new solver yet so shrug)

I just think the old name is unnecessarily obtuse though, and also kinda long -- I'd rather have the behavior not be defined by what it is "not" if possible. If you can think of a better one, though, happy to consider naming it something else.

compiler/rustc_middle/src/ty/structural_impls.rs

src/tools/clippy/clippy_utils/src/qualify_min_const_fn.rs

compiler/rustc_trait_selection/src/solve/mod.rs

compiler-errors · 2023-03-23T06:19:04Z

@rustbot ready

BoxyUwU · 2023-03-23T06:37:11Z

@bors r=BoxyUwU,lcnr
(I assume from the "apart from that and nits 👍" lcnr is fine with this PR though can always r- if misinterpreted that)

bors · 2023-03-23T06:37:13Z

📌 Commit 244cdaa has been approved by BoxyUwU,lcnr

It is now in the queue for this repository.

…xyUwU,lcnr Make alias-eq have a relation direction (and rename it to alias-relate) Emitting an "alias-eq" is too strict in some situations, since we don't always want strict equality between a projection and rigid ty. Adds a relation direction. * I could probably just reuse this [`RelationDir`](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_infer/infer/combine/enum.RelationDir.html) -- happy to uplift that struct into middle and use that instead, but I didn't feel compelled to... 🤷 * Some of the matching in `compute_alias_relate_goal` is a bit verbose -- I guess I could simplify it by using [`At::relate`](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_infer/infer/at/struct.At.html#method.relate) and mapping the relation-dir to a variance. * Alternatively, I coulld simplify things by making more helper functions on `EvalCtxt` (e.g. `EvalCtxt::relate_with_direction(T, T)` that also does the nested goal registration). No preference. r? `@lcnr` cc `@BoxyUwU` though boxy can claim it if she wants NOTE: first commit is all the changes, the second is just renaming stuff

…xyUwU,lcnr Make alias-eq have a relation direction (and rename it to alias-relate) Emitting an "alias-eq" is too strict in some situations, since we don't always want strict equality between a projection and rigid ty. Adds a relation direction. * I could probably just reuse this [`RelationDir`](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_infer/infer/combine/enum.RelationDir.html) -- happy to uplift that struct into middle and use that instead, but I didn't feel compelled to... 🤷 * Some of the matching in `compute_alias_relate_goal` is a bit verbose -- I guess I could simplify it by using [`At::relate`](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_infer/infer/at/struct.At.html#method.relate) and mapping the relation-dir to a variance. * Alternatively, I coulld simplify things by making more helper functions on `EvalCtxt` (e.g. `EvalCtxt::relate_with_direction(T, T)` that also does the nested goal registration). No preference. r? ``@lcnr`` cc ``@BoxyUwU`` though boxy can claim it if she wants NOTE: first commit is all the changes, the second is just renaming stuff

…iaskrgr Rollup of 7 pull requests Successful merges: - rust-lang#108541 (Suppress `opaque_hidden_inferred_bound` for nested RPITs) - rust-lang#109137 (resolve: Querify most cstore access methods (subset 2)) - rust-lang#109380 (add `known-bug` test for unsoundness issue) - rust-lang#109462 (Make alias-eq have a relation direction (and rename it to alias-relate)) - rust-lang#109475 (Simpler checked shifts in MIR building) - rust-lang#109504 (Stabilize `arc_into_inner` and `rc_into_inner`.) - rust-lang#109506 (make param bound vars visibly bound vars with -Zverbose) Failed merges: r? `@ghost` `@rustbot` modify labels: rollup

rustbot assigned lcnr Mar 21, 2023

lcnr reviewed Mar 22, 2023

View reviewed changes

compiler/rustc_infer/src/infer/nll_relate/mod.rs Outdated Show resolved Hide resolved

compiler/rustc_trait_selection/src/solve/mod.rs Outdated Show resolved Hide resolved

lcnr 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 Mar 22, 2023

rust-cloud-vms bot force-pushed the alias-relate branch from c1fc243 to 290080c Compare March 22, 2023 17:03

This comment has been minimized.

Sign in to view

compiler-errors force-pushed the alias-relate branch from 290080c to c54f297 Compare March 22, 2023 21:08

BoxyUwU reviewed Mar 23, 2023

View reviewed changes

compiler-errors added 4 commits March 23, 2023 05:56

Include relation direction in AliasEq predicate

5dc3fd7

Rename AliasEq -> AliasRelate

3a36a09

Printing alias-relate goals correctly

f6fd754

Remove AliasRelationDirection::Supertype

244cdaa

rust-cloud-vms bot force-pushed the alias-relate branch from c54f297 to 244cdaa Compare March 23, 2023 06:02

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 Mar 23, 2023

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 Mar 23, 2023

matthiaskrgr mentioned this pull request Mar 23, 2023

Rollup of 9 pull requests #109513

Closed

This was referenced Mar 23, 2023

Rollup of 8 pull requests #109516

Closed

Rollup of 7 pull requests #109517

Merged

bors merged commit 5d28853 into rust-lang:master Mar 23, 2023

rustbot added this to the 1.70.0 milestone Mar 23, 2023

xFrednet mentioned this pull request Mar 23, 2023

Bump rust nightly -> 2023-04-20 rust-marker/marker#123

Closed

compiler-errors deleted the alias-relate branch August 11, 2023 19:54

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Make alias-eq have a relation direction (and rename it to alias-relate) #109462

Make alias-eq have a relation direction (and rename it to alias-relate) #109462

compiler-errors commented Mar 21, 2023 •

edited

Loading

rustbot commented Mar 21, 2023

lcnr left a comment

This comment has been minimized.

BoxyUwU Mar 23, 2023

compiler-errors Mar 23, 2023

compiler-errors commented Mar 23, 2023

BoxyUwU commented Mar 23, 2023

bors commented Mar 23, 2023

Make alias-eq have a relation direction (and rename it to alias-relate) #109462

Make alias-eq have a relation direction (and rename it to alias-relate) #109462

Conversation

compiler-errors commented Mar 21, 2023 • edited Loading

rustbot commented Mar 21, 2023

lcnr left a comment

Choose a reason for hiding this comment

This comment has been minimized.

BoxyUwU Mar 23, 2023

Choose a reason for hiding this comment

compiler-errors Mar 23, 2023

Choose a reason for hiding this comment

compiler-errors commented Mar 23, 2023

BoxyUwU commented Mar 23, 2023

bors commented Mar 23, 2023

compiler-errors commented Mar 21, 2023 •

edited

Loading