Remove LEGACY_SESSION from benchmarks and back compat tests

[robert3005]

Getting closer

[0ax1]

@claude anything sus?

[github-actions]

Claude encountered an error after 3m 24s —— View job

---

I'll analyze this and get back to you.

[AdamGS]

If benchmarks look good, lets ship it?

[codspeed-hq]

Merging this PR will not alter performance

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

⚡ 4 improved benchmarks
❌ 2 regressed benchmarks
✅ 1577 untouched benchmarks

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	sum_i32_nullable_all_valid	36.5 µs	171.3 µs	-78.71%
❌	Simulation	slice_empty_vortex	310 ns	368.3 ns	-15.84%
⚡	Simulation	chunked_bool_canonical_into[(1000, 10)]	26.8 µs	16.1 µs	+66.73%
⚡	Simulation	bitwise_not_vortex_buffer_mut[128]	273.6 ns	215.3 ns	+27.1%
⚡	Simulation	bitwise_not_vortex_buffer_mut[1024]	333.9 ns	275.6 ns	+21.17%
⚡	Simulation	bitwise_not_vortex_buffer_mut[2048]	427.8 ns	369.4 ns	+15.79%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing rk/morenolegacy (a3aa8f3) with develop (71319b7)}