Remove predicate queries

[compiler-errors]

Removes:

• predicates_defined_on - Not necessary. Literally unused. Get rid of predicates_defined_on #129546
• trait_explicit_predicates_and_bounds - Not necessary technically. The side-effect of this is that we now don't remove where Self::Assoc: Bound item bounds from the where clauses of traits. This may require more trait solving to be done, but should be fine. Let's see what the fallout of this is.

r? @ghost

[compiler-errors]

@bors try @rust-timer queue

[bors]

⌛ Trying commit 8ca62f0 with merge 8104373...

[bors]

☀️ Try build successful - checks-actions
Build commit: 8104373 (8104373427fae4f23eca1b02f2790d02f929e9e2)

[rust-timer]

Finished benchmarking commit (8104373): comparison URL.

Overall result: ✅ improvements - 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)	0.4%	[0.4%, 0.4%]	1
Improvements ✅ (primary)	-0.4%	[-1.4%, -0.2%]	112
Improvements ✅ (secondary)	-0.7%	[-2.0%, -0.1%]	70
All ❌✅ (primary)	-0.4%	[-1.4%, -0.2%]	112

Max RSS (memory usage)

Results (primary -0.8%, secondary -4.0%)

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)	1.2%	[1.1%, 1.3%]	2
Regressions ❌ (secondary)	5.7%	[5.7%, 5.7%]	1
Improvements ✅ (primary)	-1.3%	[-3.3%, -0.6%]	8
Improvements ✅ (secondary)	-4.9%	[-8.5%, -2.3%]	11
All ❌✅ (primary)	-0.8%	[-3.3%, 1.3%]	10

Cycles

Results (secondary -2.4%)

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)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.4%	[-2.7%, -2.1%]	3
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 750.904s -> 750.388s (-0.07%)
Artifact size: 338.92 MiB -> 338.91 MiB (-0.00%)

[compiler-errors]

:3

r? lcnr

ill leave this to them for a vibe check when theyre back

[compiler-errors]

@bors try @rust-timer queue

[bors]

⌛ Trying commit 8ca62f0 with merge 2e9a6da...

[bors]

☀️ Try build successful - checks-actions
Build commit: 2e9a6da (2e9a6da9e514e64d0fdec2ba47c1506c388d75ba)

[rust-timer]

Finished benchmarking commit (2e9a6da): comparison URL.

Overall result: no relevant changes - 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 benchmark run did not return any relevant results for this metric.

Max RSS (memory usage)

Results (primary -2.1%, secondary 2.2%)

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)	2.2%	[1.5%, 3.0%]	2
Improvements ✅ (primary)	-2.1%	[-2.1%, -2.1%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.1%	[-2.1%, -2.1%]	1

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: 752.822s -> 753.788s (0.13%)
Artifact size: 338.90 MiB -> 338.86 MiB (-0.01%)

[compiler-errors]

OK, well no perf by itself but it removes a major source of confusion in what's required to prove vs what we can assume whatever IMO.

[lcnr]

I worry that this is a theoretical breaking change with the current cycle semantics.

Something like this probably won't cause issues, because this change is only impacting trait where-clauses.

trait Bound {}
trait Trait
where
    Self::Assoc: Bound,
{
    type Assoc;    
}

impl<T: Bound> Trait for T {
    type Assoc = T;
}
impl<T: Trait> Bound for T {}

I don't think we currently use predicates_of of the trait inside of the trait solver, so this should be alright? Would have to look at all the use-sites to make sure of that though 😅

[bors]

☔ The latest upstream changes (presumably #132027) made this pull request unmergeable. Please resolve the merge conflicts.

[lcnr]

we talked about this at some point, I think we ended up at r=me with

I don't think we currently use predicates_of of the trait inside of the trait solver, so this should be alright? Would have to look at all the use-sites to make sure of that though 😅

being resolved

[oli-obk]

Both solvers use explicit_super_predicates_of on traits.

The new solver does not call predicates_of on a trait anywhere.

The old solver

• has lots of diagnostics (error path) code that calls predicates_of on traits

• rust/compiler/rustc_trait_selection/src/traits/select/confirmation.rs

  Line 199 in f91fd0c

  let predicates = tcx.predicates_of(alias_ty.def_id).instantiate_own(tcx, alias_ty.args);

  which appears to be redundant?

• rust/compiler/rustc_trait_selection/src/traits/select/mod.rs

  Line 2830 in de94536

  cause.clone()

• normalization

  • rust/compiler/rustc_trait_selection/src/traits/normalize.rs

    Line 341 in 398fd90

    infcx.tcx.predicates_of(data.def_id).instantiate_own(infcx.tcx, data.args).map(

  • rust/compiler/rustc_trait_selection/src/traits/project.rs

    Line 516 in f91fd0c

    let predicates = tcx.predicates_of(alias_ty.def_id).instantiate(tcx, args);