chore: bump to rust 1.91

[joseph-isaacs]

This is required by packages we use

[codspeed-hq]

Merging this PR will degrade performance by 37.24%

⚡ 34 improved benchmarks
❌ 27 regressed benchmarks
✅ 1074 untouched benchmarks
⏩ 1268 skipped benchmarks¹

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	take_fsl_nullable_random[1024, 1000]	22.6 ms	18.4 ms	+22.69%
⚡	Simulation	take_fsl_nullable_random[256, 1000]	4.7 ms	3.8 ms	+25.52%
❌	Simulation	encode_varbinview[(1000, 2)]	255.6 µs	293.4 µs	-12.86%
⚡	Simulation	take_fsl_nullable_random[1024, 100]	2 ms	1.6 ms	+28.78%
⚡	Simulation	take_fsl_nullable_random[4096, 100]	8.1 ms	6.3 ms	+29.19%
⚡	Simulation	take_fsl_nullable_random[256, 100]	546.6 µs	443.6 µs	+23.22%
⚡	Simulation	take_fsl_nullable_random[4096, 1000]	90.8 ms	73.7 ms	+23.12%
⚡	Simulation	take_fsl_nullable_random[64, 1000]	1.3 ms	1.1 ms	+14.63%
⚡	Simulation	take_fsl_random[1024, 100]	2 ms	1.5 ms	+30.83%
❌	Simulation	bench_many_codes_few_values[1024]	295.2 µs	338.3 µs	-12.73%
⚡	Simulation	take_fsl_random[1024, 1000]	23 ms	18.3 ms	+25.83%
⚡	Simulation	take_fsl_random[256, 1000]	4.9 ms	3.8 ms	+28.37%
⚡	Simulation	bench_sparse_coverage[0.01]	338 µs	295.8 µs	+14.26%
⚡	Simulation	take_fsl_random[256, 100]	530.6 µs	422.6 µs	+25.55%
⚡	Simulation	take_fsl_random[4096, 1000]	92.7 ms	73.4 ms	+26.27%
⚡	Simulation	take_fsl_random[64, 100]	162.7 µs	145.7 µs	+11.66%
⚡	Simulation	take_fsl_random[4096, 100]	8.2 ms	6.3 ms	+30.74%
⚡	Simulation	take_fsl_random[64, 1000]	1.3 ms	1.1 ms	+16.29%
❌	Simulation	take_struct_simple	6.9 µs	8.3 µs	-16.46%
⚡	Simulation	old_raw_prim_test_between[i32, 32768]	390.5 µs	347.4 µs	+12.42%
...	...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

---

_{Comparing ji/bump-rust-1.91 (31ca44b) with develop (0d1a769)}

Footnotes

1. 1268 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[a10y]

I think the real fix is to disable default features: #6511

Note that lance has MSRV 1.88.0