wf-check hidden types at definition site

[aliemjay]

I believe this is the only way to fix both linked issues.
Fixes #114728
Fixes #114572

Ideally we would do WF-check after each subtyping relation, but that's cumbersome and most likely a perf hit. On the other extreme, we can do the WF-check in InferCtxt::register_hidden_type but that would break the tests in tests/ui/type-alias-impl-trait/wf-nested.rs.

This PR, as a middle ground, takes an advantage of the fact that the trait solver have limited occasions where subtyping occurs:

• PredicateKind::{Subtype,Coerce}: irrelevant in MIR typeck.
• PredicateKind::AliasRelate: fixed in this PR.
• higher-ranked equality, because it requires mutual subtyping A == B => A <: B, B<: A: we can ignore them. Why?

Given all of this we can rely on the fact that the trait solver will never infer an ill-formed type if the goal input types are well-formed.

To make sure that the trait solver input types are well-formed when invoked from MIR typeck:

• manually check every prove_predicate call in MIR typeck.
• wf-check the input types before registering AliasRelate obligations. Done in register_type_relate_obligation

r? @compiler-errors @lcnr
cc @oli-obk

[rustbot]

Some changes occurred to the core trait solver

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

[aliemjay]

@bors try

[bors]

⌛ Trying commit 83930b0 with merge b86c3b4b0ed7ac3cd1dc08a5d86c8f78b9e21603...

[rust-log-analyzer]

The job x86_64-gnu-llvm-15 failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

GITHUB_ACTION=__run_7
GITHUB_ACTIONS=true
GITHUB_ACTION_REF=
GITHUB_ACTION_REPOSITORY=
GITHUB_ACTOR=aliemjay
GITHUB_API_URL=https://api.github.com
GITHUB_BASE_REF=master
GITHUB_ENV=/home/runner/work/_temp/_runner_file_commands/set_env_607ed7a3-b068-476e-8ddd-1a2ab2a9483e
GITHUB_EVENT_NAME=pull_request
GITHUB_EVENT_NAME=pull_request
GITHUB_EVENT_PATH=/home/runner/work/_temp/_github_workflow/event.json
GITHUB_GRAPHQL_URL=https://api.github.com/graphql
GITHUB_HEAD_REF=opaque-wf-check
GITHUB_JOB=pr
GITHUB_PATH=/home/runner/work/_temp/_runner_file_commands/add_path_607ed7a3-b068-476e-8ddd-1a2ab2a9483e
GITHUB_REF=refs/pull/114933/merge
GITHUB_REF_NAME=114933/merge
GITHUB_REF_PROTECTED=false
---
GITHUB_SERVER_URL=https://github.com
GITHUB_SHA=88d523480097732deca19b48135a8f60c9ef8769
GITHUB_STATE=/home/runner/work/_temp/_runner_file_commands/save_state_607ed7a3-b068-476e-8ddd-1a2ab2a9483e
GITHUB_STEP_SUMMARY=/home/runner/work/_temp/_runner_file_commands/step_summary_607ed7a3-b068-476e-8ddd-1a2ab2a9483e
GITHUB_TRIGGERING_ACTOR=aliemjay
GITHUB_WORKFLOW_REF=rust-lang/rust/.github/workflows/ci.yml@refs/pull/114933/merge
GITHUB_WORKFLOW_SHA=88d523480097732deca19b48135a8f60c9ef8769
GITHUB_WORKSPACE=/home/runner/work/rust/rust
GOROOT_1_18_X64=/opt/hostedtoolcache/go/1.18.10/x64
---
failures:

---- [ui] tests/ui/traits/new-solver/auto-with-drop_tracking_mir.rs#pass stdout ----
normalized stderr:
error[E0271]: type mismatch resolving `[async fn body@$DIR/auto-with-drop_tracking_mir.rs:12:16: 21:2] <: impl Future<Output = ()>`
   |
LL |   async fn foo() {
   |  ________________^
LL | |
LL | |
LL | |     #[cfg(pass)]
LL | |     let x = &();
LL | |     drop(x);
LL | | }
   | |_^ types differ


error[E0271]: type mismatch resolving `[async fn body@$DIR/auto-with-drop_tracking_mir.rs:23:16: 23:18] <: impl Future<Output = ()>`
   |
LL | async fn bar() {}
   |                ^^ types differ


error: aborting due to 2 previous errors

For more information about this error, try `rustc --explain E0271`.



The actual stderr differed from the expected stderr.
Actual stderr saved to /checkout/obj/build/x86_64-unknown-linux-gnu/test/ui/traits/new-solver/auto-with-drop_tracking_mir.pass/auto-with-drop_tracking_mir.pass.stderr
To only update this specific test, also pass `--test-args traits/new-solver/auto-with-drop_tracking_mir.rs`


error in revision `pass`: 1 errors occurred comparing output.
status: exit status: 1
command: RUSTC_ICE="0" "/checkout/obj/build/x86_64-unknown-linux-gnu/stage2/bin/rustc" "/checkout/tests/ui/traits/new-solver/auto-with-drop_tracking_mir.rs" "-Zthreads=1" "-Zsimulate-remapped-rust-src-base=/rustc/FAKE_PREFIX" "-Ztranslate-remapped-path-to-local-path=no" "--sysroot" "/checkout/obj/build/x86_64-unknown-linux-gnu/stage2" "--target=x86_64-unknown-linux-gnu" "--cfg" "pass" "--error-format" "json" "--json" "future-incompat" "-Ccodegen-units=1" "-Zui-testing" "-Zdeduplicate-diagnostics=no" "-Zwrite-long-types-to-disk=no" "-Cstrip=debuginfo" "--emit" "metadata" "-C" "prefer-dynamic" "--out-dir" "/checkout/obj/build/x86_64-unknown-linux-gnu/test/ui/traits/new-solver/auto-with-drop_tracking_mir.pass" "-A" "unused" "-A" "internal_features" "-Crpath" "-Cdebuginfo=0" "-Lnative=/checkout/obj/build/x86_64-unknown-linux-gnu/native/rust-test-helpers" "-L" "/checkout/obj/build/x86_64-unknown-linux-gnu/test/ui/traits/new-solver/auto-with-drop_tracking_mir.pass/auxiliary" "-Ztrait-solver=next" "-Zdrop-tracking-mir" "--edition=2021"
--- stderr -------------------------------
--- stderr -------------------------------
error[E0271]: type mismatch resolving `[async fn body@/checkout/tests/ui/traits/new-solver/auto-with-drop_tracking_mir.rs:12:16: 21:2] <: impl Future<Output = ()>`
   |
LL |   async fn foo() {
   |  ________________^
   |  ________________^
LL | | //[pass]~^ ERROR type mismatch
LL | |     #[cfg(pass)]
LL | |     let x = &();
LL | |     drop(x);
LL | | }
   | |_^ types differ


error[E0271]: type mismatch resolving `[async fn body@/checkout/tests/ui/traits/new-solver/auto-with-drop_tracking_mir.rs:23:16: 23:18] <: impl Future<Output = ()>`
   |
LL | async fn bar() {}
   |                ^^ types differ

[bors]

☀️ Try build successful - checks-actions
Build commit: b86c3b4b0ed7ac3cd1dc08a5d86c8f78b9e21603 (b86c3b4b0ed7ac3cd1dc08a5d86c8f78b9e21603)

[compiler-errors]

@craterbot check

[craterbot]

👌 Experiment pr-114933 created and queued.
🤖 Automatically detected try build b86c3b4b0ed7ac3cd1dc08a5d86c8f78b9e21603
🔍 You can check out the queue and this experiment's details.

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[craterbot]

🚧 Experiment pr-114933 is now running

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[aliemjay]

@craterbot check crates=https://crater-reports.s3.amazonaws.com/pr-114933/retry-regressed-list.txt

[craterbot]

👌 Experiment pr-114933-1 created and queued.
🤖 Automatically detected try build b86c3b4b0ed7ac3cd1dc08a5d86c8f78b9e21603
🔍 You can check out the queue and this experiment's details.

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[craterbot]

🎉 Experiment pr-114933 is completed!
📊 27 regressed and 1 fixed (349709 total)
📰 Open the full report.

⚠️ If you notice any spurious failure please add them to the blacklist!
ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[craterbot]

🚧 Experiment pr-114933-1 is now running

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[craterbot]

🎉 Experiment pr-114933-1 is completed!
📊 6 regressed and 0 fixed (27 total)
📰 Open the full report.

⚠️ If you notice any spurious failure please add them to the blacklist!
ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[aliemjay]

@rust-timer build b86c3b4b0ed7ac3cd1dc08a5d86c8f78b9e21603

[rust-timer]

Finished benchmarking commit (b86c3b4b0ed7ac3cd1dc08a5d86c8f78b9e21603): comparison URL.

Overall result: ❌ regressions - no action needed

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-perf -perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	1.0%	[0.9%, 1.1%]	6
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	4.3%	[4.1%, 4.4%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Cycles

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

Binary size

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

Bootstrap: 632.38s -> 635.628s (0.51%)
Artifact size: 347.07 MiB -> 347.09 MiB (0.00%)

[oli-obk]

this regresses

• axum-starter-0.8.0
• rs-matter-0.1.0
• schemat-0.1.17

All of them have updates in the last month and at least the first two seem easily fixable.

[oli-obk]

This PR fixes a soundness hole, where we were failing to check that hidden types are actually well formed wrt to lifetimes. From a not-well formed hidden type you can easily satisfy trait bounds that make no sense and thus allow treating any lifetime as another lifetime that lives longer.

Considering that the implementation does nothing but add well-formedness predicates, which should always be sound and at worst a performance issue, I do not believe there is any possible risk with this PR, beyond causing breakage for benign code that just forgot some bounds somewhere. Crater showed 3 such cases, which we should all fix before landing this.

@rfcbot fcp merge

[rfcbot]

Team member @oli-obk has proposed to merge this. The next step is review by the rest of the tagged team members:

• @BoxyUwU
• @aliemjay
• @compiler-errors
• @jackh726
• @lcnr
• @nikomatsakis
• @oli-obk
• @spastorino

Concerns:

• we should open PRs against the crates that will break (wf-check hidden types at definition site #114933 (comment))
• why prefer this over stricter hidden type wf-check #115008 (wf-check hidden types at definition site #114933 (comment))

Once a majority of reviewers approve (and at most 2 approvals are outstanding), this will enter its final comment period. If you spot a major issue that hasn't been raised at any point in this process, please speak up!

See this document for info about what commands tagged team members can give me.

[oli-obk]

@rfcbot concern we should open PRs against the crates that will break

[jackh726]

Can we get specific minimal examples of the crates that are breaking? Are they potential soundness issues?

[lcnr]

@rfcbot concern why prefer this over #115008

We currently have 3 different PRs up to fix this bug, this one, #115008 and #114740, I consider the approach of this PR to be fairly fragile and would prefer something closer to #115008.

[oli-obk]

Closing in favor of #115008, will explain the differences and reasons on that PR when starting the FCP