Avoid an intermediate vec when ref-decoding to &'tcx [T]

[Zalathar]

For !needs_drop types, this should allow decoding directly into the arena-allocated slice.

For needs_drop types, the arena already collects into a SmallVec, so this avoids an intermediate conversion from Vec to SmallVec.

---

I was planning some cleanups to RefDecodable impls (#162287), and along the way I noticed that some hand-written impls for slices manage to avoid the overhead of decoding to an intermediate Vec, by instead reading a usize length and then decoding that many items directly. This matches the format that encoding a vec/slice or decoding to Vec would use.

This PR therefore applies that same optimization to the impls generated by impl_ref_decodable_into_arena!.

I extracted this change into its own small PR since it has perf effects.

r? nnethercote (or compiler)

[Zalathar]

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: 629cf3b (629cf3bf6d1771b98aa38784f53c8969e5187bb4)
Base parent: f248f40 (f248f4038796913873f11ca65b1b901e311c8dae)

[rust-timer]

Finished benchmarking commit (629cf3b): comparison URL.

Overall result: no relevant changes - 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

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

Max RSS (memory usage)

Results (secondary 7.4%)

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)	7.4%	[7.4%, 7.4%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Cycles

Results (secondary 3.9%)

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)	3.9%	[3.9%, 3.9%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 476.686s -> 477.652s (0.20%)
Artifact size: 403.40 MiB -> 403.46 MiB (0.01%)

[Zalathar]

Hmm, curious that this on its own didn't produce any measurable perf effect.

[Zalathar]

Based on another perf run of #162287, the changes I saw were noise, so I’ll close this PR and keep it as part of the larger cleanup PR.