interpret region vars as universals in implied bounds query

[aliemjay]

Implied bounds query is different to other canonical trait queries in that we should treat all region variables as universals. This fixes #106569.

To illustrate the problem in #106569:

• query: implied_outlives_bounds(Canonical<Equal<'^1, '^2>)
• instantiate the query: implied_outlives_bounds(Equal<'?1, '?2>)
• unify '?1 == '?2 during normalization.
• query response: ['?1: '?2, '?2: ?1] (as expected)
• when canonicalizing the query response we resolve unified region vars opportunistically. The query response is now: ['?1: '?1, '?1: '?1]

This PR mitigates this problem by mapping every existential region to a unique universal one before executing the query.

cc @lcnr
r? types

[lcnr]

this is definitely complex and feels intuitively wrong to me¹, will try to look at this at some point this week.

r? @lcnr

Footnotes

1. which doesn't mean that it necessarily is ↩

[aliemjay]

@bors try @rust-timer queue

[bors]

@aliemjay: 🔑 Insufficient privileges: not in try users

[oli-obk]

@bors try @rust-timer queue

[bors]

⌛ Trying commit 4be4829 with merge 6a52a02ebbfeb6f8adfd1aa63e64a290f201d98c...

[bors]

☀️ Try build successful - checks-actions
Build commit: 6a52a02ebbfeb6f8adfd1aa63e64a290f201d98c (6a52a02ebbfeb6f8adfd1aa63e64a290f201d98c)

[bors]

☀️ Try build successful - checks-actions
Build commit: 6a52a02ebbfeb6f8adfd1aa63e64a290f201d98c (6a52a02ebbfeb6f8adfd1aa63e64a290f201d98c)

[rust-timer]

Finished benchmarking commit (6a52a02ebbfeb6f8adfd1aa63e64a290f201d98c): 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

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)	1.6%	[0.4%, 3.0%]	7
Regressions ❌ (secondary)	2.6%	[0.9%, 4.3%]	57
Improvements ✅ (primary)	-1.1%	[-1.6%, -0.7%]	2
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	1.0%	[-1.6%, 3.0%]	9

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)	0.4%	[0.4%, 0.4%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.4%	[0.4%, 0.4%]	1

[lcnr]

we talked about this in sync at some point 😅 this is missing an explanatory comment, apart from that r=me

[aliemjay]

blocked on #109763. I touches the same and and I want to prioritize #109763.

[Dylan-DPC]

@aliemjay what's the status of this? this was blocked on a pr that's now closed but it was close to being approved

[Dylan-DPC]

Closing this as inactive. Feel free to reöpen this pr or create a new pr if you get the time to work on this. Thanks