Add all RPITITs when augmenting param-env with GAT bounds in `check_type_bounds` #117131

compiler-errors · 2023-10-24T16:42:32Z

When checking that associated type definitions actually satisfy their associated type bounds in check_type_bounds, we construct a "normalize_param_env" which adds a projection predicate that allows us to assume that we can project the GAT to the definition we're checking. For example, in:

type Foo {
  type Bar: Display = i32;
}

We would add <Self as Foo>::Bar = i32 as a projection predicate when checking that i32: Display holds.

That normalize_param_env was, for some reason, only being used to normalize the predicate before it was registered. This is sketchy, because a nested obligation may require the GAT bound to hold, and also the projection cache is broken and doesn't differentiate projection cache keys that differ by param-envs 😿.

This normalize_param_env is also not sufficient when we have nested RPITITs and default trait methods, since we need to be able to assume we can normalize both the RPITIT and all of its child RPITITs to sufficiently prove all of its bounds. This is the cause of #117104, which only starts to fail for RPITITs that are nested 3 and above due to the projection-cache bug above.¹

First fix

Use the normalize_param_env everywhere in check_type_bounds. This is reflected in a test I've constructed that fixes a GAT-only failure.

Second fix

For RPITITs, install projection predicates for each RPITIT in the same function in check_type_bounds. This fixes #117104.

not sure who to request, so...
r? @lcnr hehe feel free to reassign :3

The projection cache bug specifically occurs because we try normalizing the assumed_wf_types with the non-normalization param-env. This causes us to insert a projection cache entry that keeps the outermost RPITIT rigid, and it trivially satisifes all its own bounds. Super sketchy!² ↩
I haven't actually gone and fixed the projection cache bug because it's only marginally related, but I could, and it should no longer be triggered here. ↩

axos88 · 2023-10-27T21:23:20Z

Is there a (relatively) easy way to build a compiler from this PR before it gets merged? This bug is blocking me.

lcnr · 2023-11-02T10:24:42Z

@bors r+ rollup

bors · 2023-11-02T10:24:46Z

📌 Commit bfd836b has been approved by lcnr

It is now in the queue for this repository.

lcnr · 2023-11-02T10:25:12Z

@bors rollup=never

bors · 2023-11-02T12:20:45Z

⌛ Testing commit bfd836b with merge 5626f49...

Add all RPITITs when augmenting param-env with GAT bounds in `check_type_bounds` When checking that associated type definitions actually satisfy their associated type bounds in `check_type_bounds`, we construct a "`normalize_param_env`" which adds a projection predicate that allows us to assume that we can project the GAT to the definition we're checking. For example, in: ```rust type Foo { type Bar: Display = i32; } ``` We would add `<Self as Foo>::Bar = i32` as a projection predicate when checking that `i32: Display` holds. That `normalize_param_env` was, for some reason, only being used to normalize the predicate before it was registered. This is sketchy, because a nested obligation may require the GAT bound to hold, and also the projection cache is broken and doesn't differentiate projection cache keys that differ by param-envs 😿. This `normalize_param_env` is also not sufficient when we have nested RPITITs and default trait methods, since we need to be able to assume we can normalize both the RPITIT and all of its child RPITITs to sufficiently prove all of its bounds. This is the cause of rust-lang#117104, which only starts to fail for RPITITs that are nested 3 and above due to the projection-cache bug above.[^1] ## First fix Use the `normalize_param_env` everywhere in `check_type_bounds`. This is reflected in a test I've constructed that fixes a GAT-only failure. ## Second fix For RPITITs, install projection predicates for each RPITIT in the same function in `check_type_bounds`. This fixes rust-lang#117104. not sure who to request, so... r? `@lcnr` hehe feel free to reassign :3 [^1]: The projection cache bug specifically occurs because we try normalizing the `assumed_wf_types` with the non-normalization param-env. This causes us to insert a projection cache entry that keeps the outermost RPITIT rigid, and it trivially satisifes all its own bounds. Super sketchy![^2] [^2]: I haven't actually gone and fixed the projection cache bug because it's only marginally related, but I could, and it should no longer be triggered here.

bors · 2023-11-02T12:36:10Z

💔 Test failed - checks-actions

…pe_bounds

compiler-errors · 2023-11-02T20:48:08Z

@bors r=lcnr

bors · 2023-11-02T20:48:10Z

📌 Commit dd571e4 has been approved by lcnr

It is now in the queue for this repository.

bors · 2023-11-03T00:02:48Z

⌛ Testing commit dd571e4 with merge 2520ca8...

bors · 2023-11-03T02:03:59Z

☀️ Test successful - checks-actions
Approved by: lcnr
Pushing 2520ca8 to master...

rust-timer · 2023-11-03T03:31:20Z

Finished benchmarking commit (2520ca8): comparison URL.

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

Next Steps: If you can justify the regressions found in this perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please open an issue or create a new PR that fixes the regressions, add a comment linking to the newly created issue or PR, and then add the perf-regression-triaged label to this PR.

@rustbot label: +perf-regression
cc @rust-lang/wg-compiler-performance

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)	32.3%	[0.3%, 99.8%]	19
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.4%	[-1.7%, -1.0%]	6
All ❌✅ (primary)	32.3%	[0.3%, 99.8%]	19

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)	16.1%	[8.0%, 24.4%]	13
Regressions ❌ (secondary)	1.7%	[0.8%, 2.6%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	16.1%	[8.0%, 24.4%]	13

Cycles

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)	46.4%	[7.5%, 105.5%]	13
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	46.4%	[7.5%, 105.5%]	13

Binary size

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

Bootstrap: 635.107s -> 636.745s (0.26%)
Artifact size: 304.48 MiB -> 304.48 MiB (-0.00%)

82: Automated pull from upstream `master` r=tshepang a=github-actions[bot] This PR pulls the following changes from the upstream repository: * rust-lang/rust#117313 * rust-lang/rust#117131 * rust-lang/rust#117134 * rust-lang/rust#117471 * rust-lang/rust#117521 * rust-lang/rust#117513 * rust-lang/rust#117512 * rust-lang/rust#117509 * rust-lang/rust#117495 * rust-lang/rust#117394 * rust-lang/rust#117466 * rust-lang/rust#117204 * rust-lang/rust#117386 * rust-lang/rust#117506 Co-authored-by: Nicholas Nethercote <n.nethercote@gmail.com> Co-authored-by: roblabla <unfiltered@roblab.la> Co-authored-by: Michael Goulet <michael@errs.io> Co-authored-by: massivebird <gdrakemail@gmail.com> Co-authored-by: bors <bors@rust-lang.org> Co-authored-by: Zalathar <Zalathar@users.noreply.github.com> Co-authored-by: lcnr <rust@lcnr.de> Co-authored-by: Joshua Liebow-Feeser <joshlf@users.noreply.github.com> Co-authored-by: Matthias Krüger <matthias.krueger@famsik.de>

lcnr · 2023-11-03T09:18:40Z

alright... that's bad 😅 how tf does eval_to_allocation_raw take that much longer because of this PR 😅 I think we should either solve these regressions or revert before the beta cutoff

Only instantiate binder during dyn's built-in trait candidate probe once See UI test for demonstration of the issue. This was "caused" by rust-lang#117131, but only because we're using the `normalize_param_env` (which has been augmented with a projection clause used to normalize GATs) which features non-lifetime bound vars in it. Fixes rust-lang#117602 technically, though that's also fixed by rust-lang#117542. r? types

pnkfelix · 2023-11-07T19:12:24Z

the regression to bitmaps (to 13 variants, ranging from 10.8% to 99.8%) appears persistent (non-noise)
it carries over to cycle-count and wall-clock time.
compiler-errors has been investigating, and it looks like PR Only use normalize_param_env when normalizing predicate in check_item_bounds #117542 will address it.
marking as triaged

@rustbot label: +perf-regression-triaged

rustbot assigned lcnr Oct 24, 2023

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 Oct 24, 2023

rust-cloud-vms bot force-pushed the projection-oops branch from 3d251d8 to d40fd3d Compare October 24, 2023 16:42

This comment has been minimized.

Sign in to view

rust-cloud-vms bot force-pushed the projection-oops branch 2 times, most recently from ca6f45c to bfd836b Compare October 25, 2023 02:30

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 Nov 2, 2023

This comment has been minimized.

Sign in to view

bors added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Nov 2, 2023

rust-cloud-vms bot force-pushed the projection-oops branch from bfd836b to b0ee801 Compare November 2, 2023 20:45

compiler-errors added 2 commits November 2, 2023 20:47

Use the normalizing param-env always in check_type_bounds

bb74d7e

Add all RPITITs when augmenting param-env with GAT bounds in check_ty…

dd571e4

…pe_bounds

rust-cloud-vms bot force-pushed the projection-oops branch from b0ee801 to dd571e4 Compare November 2, 2023 20:47

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 Nov 2, 2023

bors added the merged-by-bors This PR was explicitly merged by bors. label Nov 3, 2023

bors merged commit 2520ca8 into rust-lang:master Nov 3, 2023
12 checks passed

rustbot added this to the 1.75.0 milestone Nov 3, 2023

rustbot added the perf-regression Performance regression. label Nov 3, 2023

github-actions bot mentioned this pull request Nov 3, 2023

Automated pull from upstream master ferrocene/ferrocene#82

Merged

This was referenced Nov 3, 2023

Only use normalize_param_env when normalizing predicate in check_item_bounds #117542

Merged

Scrub param-env when evaluating constant with no generic args #117543

Closed

HTGAzureX1212 mentioned this pull request Nov 5, 2023

Recent nightly causes [E0308]: mismatched types #117598

Closed

lqd mentioned this pull request Nov 5, 2023

Nightly 65e297d1e 2023-11-03 fails to infer auto traits #117602

Closed

compiler-errors mentioned this pull request Nov 5, 2023

Only instantiate binder during dyn's built-in trait candidate probe once #117610

Merged

Noratrieb mentioned this pull request Nov 7, 2023

regression: error compiling sqlx #117660

Closed

lqd mentioned this pull request Nov 7, 2023

New "lifetime bound not satisfied" error on associated type #117664

Closed

pnkfelix mentioned this pull request Nov 7, 2023

Add track_caller to transmute_copy #117510

Merged

rustbot added the perf-regression-triaged The performance regression has been triaged. label Nov 7, 2023

compiler-errors mentioned this pull request Apr 21, 2024

fix normalizing in different ParamEnvs with the same InferCtxt #124203

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add all RPITITs when augmenting param-env with GAT bounds in `check_type_bounds` #117131

Add all RPITITs when augmenting param-env with GAT bounds in `check_type_bounds` #117131

compiler-errors commented Oct 24, 2023

This comment has been minimized.

axos88 commented Oct 27, 2023

lcnr commented Nov 2, 2023

bors commented Nov 2, 2023

lcnr commented Nov 2, 2023

bors commented Nov 2, 2023

This comment has been minimized.

bors commented Nov 2, 2023

compiler-errors commented Nov 2, 2023

bors commented Nov 2, 2023

bors commented Nov 3, 2023

bors commented Nov 3, 2023

rust-timer commented Nov 3, 2023

lcnr commented Nov 3, 2023

pnkfelix commented Nov 7, 2023

Add all RPITITs when augmenting param-env with GAT bounds in check_type_bounds #117131

Add all RPITITs when augmenting param-env with GAT bounds in check_type_bounds #117131

Conversation

compiler-errors commented Oct 24, 2023

First fix

Second fix

Footnotes

This comment has been minimized.

axos88 commented Oct 27, 2023

lcnr commented Nov 2, 2023

bors commented Nov 2, 2023

lcnr commented Nov 2, 2023

bors commented Nov 2, 2023

This comment has been minimized.

bors commented Nov 2, 2023

compiler-errors commented Nov 2, 2023

bors commented Nov 2, 2023

bors commented Nov 3, 2023

bors commented Nov 3, 2023

rust-timer commented Nov 3, 2023

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

Instruction count

Max RSS (memory usage)

Cycles

Binary size

lcnr commented Nov 3, 2023

pnkfelix commented Nov 7, 2023

Add all RPITITs when augmenting param-env with GAT bounds in `check_type_bounds` #117131

Add all RPITITs when augmenting param-env with GAT bounds in `check_type_bounds` #117131