perf: faster WarmAddresses short precompile bits lookup

[borngraced]

is_warm/is_cold (via the short precompile │ bitset) is on a very hot path. This optimization removes the bitvec + Arc overhead by using a fixed size array [u8; 38] with direct bit manipulation, matching the technique from #2618.

This is a rough estimate microbenchmark numbers from my machine.. can be faster on yours.

Benchmark

Implementation	Time per op	Ops/sec	Speedup
Fixed Array + Manual Bits	9.15 ns	109.3M	10.1x faster
BitVec (legacy)	92.62 ns	10.8M	baseline

Pattern	Fixed Array	BitVec	Speedup
Sequential	8.84 ns	92.24 ns	10.4x
Random	8.43 ns	90.83 ns	10.8x

Also got rid of the bitvec dep from context crate

[codspeed-hq]

Merging this PR will improve performance by 4.02%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 1 improved benchmark
✅ 180 untouched benchmarks
⏩ 1 skipped benchmark¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	EXTCODECOPY_50	51.7 µs	49.7 µs	+4.02%

Tip

Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing borngraced:faster-warm-addresses-short-precompile-bits-lookup (f9e7ff3) with main (cfcf038)}

Footnotes

1. 1 benchmark was skipped, so the baseline result was used instead. If it was deleted from the codebase, click here and archive it to remove it from the performance reports. ↩