mir_build: Clearly distinguish or/refutable/irrefutable patterns during match lowering

[Zalathar]

This PR has two related parts.

• The first commit divides the various fields of InterPat into three different InterPatKind variants, which clearly distinguish between or-pattern nodes, refutable pattern nodes, and irrefutable pattern nodes. Such a split was suggested during review of mir_build: Add an extra intermediate step in MIR building for patterns  #155144, but at the time I wasn't sure if it was the right direction. Now that I've thought more about how to make it work nicely with InterPat node building, I'm more comfortable with the explicit split.

• As a follow-on, the second commit then splits MatchPairTree into two different MatchPairKind variants: one for or-pattern nodes, and one for “testable” nodes.

In both cases, dividing the outer struct into multiple kinds makes it easier to understand and uphold field invariants, since previously-optional fields can be made mandatory or absent.

There should be no change to compiler output.

---

• The changes to MatchPairTree can be seen as an alternative to mir_build: Split MatchPairTree into testable and or-pattern variants #158551. Instead of splitting or-patterns and testable-patterns into separate lists, they still occupy the same list but are distinguished by kind.

---

r? Nadrieril

[rustbot]

Some changes occurred in match lowering

cc @Nadrieril

[rustbot]

Nadrieril is currently at their maximum review capacity.
They may take a while to respond.

[rustbot]

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[Zalathar]

This PR isn't motivated by perf, but I'm curious to see if it has any measurable perf effect.

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: fcfa9a7 (fcfa9a7ac1652c381facd5e795b16ffc8808607d)
Base parent: f73951d (f73951df0a5566d94d13b7954acd9f4ab1fa3734)

[rust-timer]

Finished benchmarking commit (fcfa9a7): comparison URL.

Overall result: ❌ regressions - no action needed

Benchmarking means the PR may be perf-sensitive. Consider adding rollup=never if this change is not fit for rolling up.

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

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.1%	[0.1%, 0.1%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary -1.5%, secondary -2.0%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	1.5%	[0.8%, 2.1%]	2
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-3.5%	[-6.0%, -2.1%]	3
Improvements ✅ (secondary)	-2.0%	[-2.0%, -2.0%]	1
All ❌✅ (primary)	-1.5%	[-6.0%, 2.1%]	5

Cycles

This perf run didn't have relevant results for this metric.

Binary size

Results (primary -0.0%, secondary -0.0%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.0%	[-0.0%, -0.0%]	11
Improvements ✅ (secondary)	-0.0%	[-0.1%, -0.0%]	5
All ❌✅ (primary)	-0.0%	[-0.0%, -0.0%]	11

Bootstrap: 460.81s -> 462.278s (0.32%)
Artifact size: 398.76 MiB -> 398.61 MiB (-0.04%)

[Zalathar]

A tiny regression in match-stress is potentially real, but not worth worrying about.