Add Output = expected type trait obligation for known binary operators #96482

willcrichton · 2022-04-27T17:03:21Z

This PR is a follow-on to #94034 that addresses #96442. That is, after replacing the trait-suggestion logic in op.rs with a more generic path that analyzes a general set of Obligations, then we lost some specificity in the suggestions where the bounds on the associated type Output= would not get suggested.

This PR fixes this issue by changing FnCtxt::construct_obligation_for_trait to include a new ProjectionPredicate obligation for binary operators that obliges that Output is the same as the expected type of the expression. Additionally, to get the expected type of the expression, this PR threads the Expectation<'tcx> structure throughout several functions.

See src/test/ui/generic-associated-types/missing-bounds.stderr for an example of how this works.

One side effect of this change is it causes type-check failures with binops to include additional information. Specifically, many now say

error: type mismatch resolving `<Lhs as TheBinop>::Output == ExpectedTy`

It's up for discussion whether this added context is worth it to the user.

r? @estebank

compiler-errors · 2022-04-27T17:08:31Z

@bors try @rust-timer queue

rust-timer · 2022-04-27T17:08:32Z

Awaiting bors try build completion.

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

bors · 2022-04-27T17:08:46Z

⌛ Trying commit 6fe3b1debd3b44413c657a0b1dba37d553d7a526 with merge 2142253261d42f925cbaf34fe261e47feea3b67b...

willcrichton · 2022-04-27T17:34:01Z

(ah I missed some tests, will fix those later today)

compiler-errors · 2022-04-27T17:44:28Z

retrying perf run, pushes will cause it to cancel

@bors try @rust-timer queue

rust-timer · 2022-04-27T17:44:30Z

Awaiting bors try build completion.

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

bors · 2022-04-27T17:44:37Z

⌛ Trying commit e9ecae53d06562c178dfb3a81d298f5c167639b9 with merge f07f0f43ceab58a4ed0051bac441d2d07d6571e2...

compiler-errors · 2022-04-27T17:45:50Z

compiler/rustc_typeck/src/check/method/mod.rs

+                obligations.push(traits::Obligation::new(
+                    cause.clone(),
+                    self.param_env,
+                    ty::Binder::dummy(ty::PredicateKind::Projection(ProjectionPredicate {


My only worry is that registering this projection obligation will cause inference to break in certain places, since it asserts type equality as opposed to coercibility.

compiler-errors · 2022-04-27T17:47:35Z

compiler/rustc_typeck/src/check/method/mod.rs

@@ -511,6 +516,31 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> {
        };
        obligations.extend(traits::predicates_for_generics(cause.clone(), self.param_env, bounds));


It might be possible to stash the type from the expectation in the BinOp cause-code then edit the code that suggests "add std::ops::Mul" to add the associated type there, in case inference is broken by the current change (or if the perf hit is too massive). I'll try to think about if either is a really valid worry, though.

in case inference is broken by the current change

I guess not necessarily, since we're only applying this when we have Expectation::ExpectHasType, not the other ones which are there to guide coercion... (I think)

compiler-errors · 2022-04-27T17:49:09Z

compiler/rustc_typeck/src/check/method/mod.rs

@@ -511,6 +516,31 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> {
        };
        obligations.extend(traits::predicates_for_generics(cause.clone(), self.param_env, bounds));

+        // If we're producing obligations for an operator trait, and it has an `Output`
+        // associated type, then we add an `Output=expected` obligation.
+        if let (true, Expectation::ExpectHasType(output_ty)) = (is_op, expected) {


nits: let chains can make this more readable, and Expectation::only_has_type might be useful instead of matching on the variant itself.

bors · 2022-04-27T19:06:05Z

☀️ Try build successful - checks-actions
Build commit: f07f0f43ceab58a4ed0051bac441d2d07d6571e2 (f07f0f43ceab58a4ed0051bac441d2d07d6571e2)

rust-timer · 2022-04-27T19:06:07Z

Queued f07f0f43ceab58a4ed0051bac441d2d07d6571e2 with parent ff18038, future comparison URL.

rust-timer · 2022-04-27T21:31:12Z

Finished benchmarking commit (f07f0f43ceab58a4ed0051bac441d2d07d6571e2): comparison url.

Summary: This benchmark run did not return any relevant results.

If you disagree with this performance assessment, please file an issue in rust-lang/rustc-perf.

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR may lead to changes in compiler perf.

@bors rollup=never
@rustbot label: +S-waiting-on-review -S-waiting-on-perf -perf-regression

estebank · 2022-04-28T00:56:14Z

I'm not too happy with the increase in verbosity, but the perf impact isn't massive and we might be able to work things out so that we don't regress as much in the output.

error[E0271]: type mismatch resolving `<! as std::ops::Not>::Output == bool`
    |
    |
239 |     if !return () {}
    |        ^^^^^^^^^^ expected `!`, found `bool`
    = note: expected type `!`
               found type `bool`

This is a problem, though :(

I think this suggestion might be a more straightforward approach to get the same effect we're looking for.

willcrichton · 2022-04-28T06:54:50Z

Yeah I think my original strategy won't work if it's changing type-checking results. I tried what @compiler-errors suggested of carrying additional data in the obligation cause, and passing that through to the suggestion code. Take a look and let me know if the implementation seems reasonable. It fixes the specific issue in missing-bounds.rs while not breaking anything else.

The hackiest part is the code that generates and appends the string <Type=Expected> to the trait constraint. I'm not sure if there's a better way to do that.

bors · 2022-05-18T10:27:24Z

☔ The latest upstream changes (presumably #97135) made this pull request unmergeable. Please resolve the merge conflicts.

estebank

r=me after rebase (ping me to approve)

estebank · 2022-05-18T18:11:22Z

compiler/rustc_typeck/src/check/op.rs

+                        if !errors.is_empty() {
+                            for error in errors {
+                                if let Some(trait_pred) =
+                                    error.obligation.predicate.to_opt_poly_trait_pred()
+                                {


This is to avoid the Vec allocation, right? Good idea.

estebank · 2022-05-18T18:12:22Z

compiler/rustc_typeck/src/check/method/mod.rs

@@ -504,6 +525,7 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> {
                    rhs_span: opt_input_expr.map(|expr| expr.span),
                    is_lit: opt_input_expr
                        .map_or(false, |expr| matches!(expr.kind, hir::ExprKind::Lit(_))),
+                    output_pred: None,


Not necessary for this PR: I wonder if we can provide the predicate somehow here 🤔

estebank · 2022-05-18T18:14:41Z

Apologies for the delay in reviewing. If you can squash the commits, that would also be useful.

estebank · 2022-06-02T17:35:18Z

Ping :)

willcrichton · 2022-06-02T18:01:08Z

Ah apologies for the delay in the response to the review! Too busy graduating. I will get to this by the weekend.

estebank · 2022-06-02T20:19:41Z

@willcrichton no hurries! Congratulations!

JohnCSimon · 2022-07-03T01:50:16Z

ping from triage:
@willcrichton -

estebank commented on May 18
Apologies for the delay in reviewing. If you can squash the commits, that would also be useful.

FYI: when a PR is ready for review, send a message containing
@rustbot ready to switch to S-waiting-on-review so the PR is in the reviewer's backlog.

willcrichton · 2022-07-06T20:31:39Z

Sorry again for the delays. I just rebased, and found a new regression. Whatever code is producing the suggestion is creating an incorrect syntax combining the associated type bound with an existing trait parameter. See: https://github.com/rust-lang/rust/pull/96482/files#diff-b4e22a434e69c155966cdab3f3b6ae44922478447624ba038055ca32f1b3a50f

I will try to debug this soon.

bors · 2022-07-12T03:32:58Z

☔ The latest upstream changes (presumably #99165) made this pull request unmergeable. Please resolve the merge conflicts.

estebank · 2022-07-14T15:35:58Z

src/test/ui/suggestions/issue-97677.stderr

+LL | fn add_ten<N: std::ops::Add<i32><Output=N>>(n: N) -> N {
+   |             ++++++++++++++++++++++++++++++


If you can't figure this one out, let's clean up for merge and open a follow up ticket.

I figured out the issue. I added a fix, although it duplicates some code as you can see in the commit and the FIXME. If you want me to figure out a way to factor out the duplicated code I can do that, otherwise the PR is ready I think.

…ise trait obligations for binops.

…it bounds

estebank · 2022-07-16T02:40:18Z

@bors r+

bors · 2022-07-16T02:40:20Z

📌 Commit 2f15dfa has been approved by estebank

It is now in the queue for this repository.

bors · 2022-07-16T11:36:21Z

⌛ Testing commit 2f15dfa with merge d695a49...

bors · 2022-07-16T14:05:46Z

☀️ Test successful - checks-actions
Approved by: estebank
Pushing d695a49 to master...

rust-timer · 2022-07-16T15:22:51Z

Finished benchmarking commit (d695a49): comparison url.

Instruction count

This benchmark run did not return any relevant results for this metric.

Max RSS (memory usage)

Results

Primary benchmarks: 😿 relevant regression found
Secondary benchmarks: 😿 relevant regressions found

	mean¹	max	count²
Regressions 😿 (primary)	2.9%	2.9%	1
Regressions 😿 (secondary)	2.9%	4.3%	3
Improvements 🎉 (primary)	N/A	N/A	0
Improvements 🎉 (secondary)	N/A	N/A	0
All 😿🎉 (primary)	2.9%	2.9%	1

Cycles

Results

Primary benchmarks: no relevant changes found
Secondary benchmarks: 😿 relevant regression found

	mean¹	max	count²
Regressions 😿 (primary)	N/A	N/A	0
Regressions 😿 (secondary)	5.8%	5.8%	1
Improvements 🎉 (primary)	N/A	N/A	0
Improvements 🎉 (secondary)	N/A	N/A	0
All 😿🎉 (primary)	N/A	N/A	0

If you disagree with this performance assessment, please file an issue in rust-lang/rustc-perf.

@rustbot label: -perf-regression

the arithmetic mean of the percent change ↩ ↩²
number of relevant changes ↩ ↩²

rust-highfive assigned estebank Apr 27, 2022

rustbot added the T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. label Apr 27, 2022

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Apr 27, 2022

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Apr 27, 2022

compiler-errors reviewed Apr 27, 2022

View reviewed changes

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 Apr 27, 2022

estebank approved these changes May 18, 2022

View reviewed changes

estebank 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 Jun 2, 2022

willcrichton force-pushed the fix-trait-suggestion-for-binops branch from c40c425 to b61f8ba Compare July 6, 2022 20:29

estebank reviewed Jul 14, 2022

View reviewed changes

willcrichton added 2 commits July 15, 2022 18:06

Propagate Expectation around binop typeck code to construct more prec…

e5bb7d8

…ise trait obligations for binops.

Fix suggestion regression with incorrect syntactic combination of tra…

2f15dfa

…it bounds

willcrichton force-pushed the fix-trait-suggestion-for-binops branch from ba5f969 to 2f15dfa Compare July 16, 2022 01:06

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-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jul 16, 2022

bors added the merged-by-bors This PR was explicitly merged by bors. label Jul 16, 2022

bors merged commit d695a49 into rust-lang:master Jul 16, 2022

rustbot added this to the 1.64.0 milestone Jul 16, 2022

		@@ -511,6 +516,31 @@ impl<'a, 'tcx> FnCtxt<'a, 'tcx> {
		};
		obligations.extend(traits::predicates_for_generics(cause.clone(), self.param_env, bounds));

		LL \| fn add_ten<N: std::ops::Add<i32><Output=N>>(n: N) -> N {
		\| ++++++++++++++++++++++++++++++

Add Output = expected type trait obligation for known binary operators #96482

Add Output = expected type trait obligation for known binary operators #96482

Conversation

willcrichton commented Apr 27, 2022 • edited by estebank Loading

compiler-errors commented Apr 27, 2022

rust-timer commented Apr 27, 2022

bors commented Apr 27, 2022

willcrichton commented Apr 27, 2022

compiler-errors commented Apr 27, 2022

rust-timer commented Apr 27, 2022

bors commented Apr 27, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

This comment has been minimized.

bors commented Apr 27, 2022

rust-timer commented Apr 27, 2022

rust-timer commented Apr 27, 2022

estebank commented Apr 28, 2022

willcrichton commented Apr 28, 2022 • edited Loading

bors commented May 18, 2022

estebank left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

estebank commented May 18, 2022

estebank commented Jun 2, 2022

willcrichton commented Jun 2, 2022

estebank commented Jun 2, 2022

JohnCSimon commented Jul 3, 2022

willcrichton commented Jul 6, 2022

bors commented Jul 12, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

estebank commented Jul 16, 2022

bors commented Jul 16, 2022

bors commented Jul 16, 2022

bors commented Jul 16, 2022

rust-timer commented Jul 16, 2022

Footnotes

willcrichton commented Apr 27, 2022 •

edited by estebank

Loading

willcrichton commented Apr 28, 2022 •

edited

Loading