wip: Remove lifetime from scalars

[kylebarron]

Using Arrow's slicing is almost certainly easier than working with the raw buffers a second time

• Still need to fix some tests
• Still need to run the benchmarks to ensure there's no obvious regression

Welp... it's quite a lot slower to iterate over values! In this case the simplicity isn't worth the performance overhead

     Running benches/area.rs (target/release/deps/area-f218a6feb2e2143b)
Gnuplot not found, using plotters backend
area                    time:   [195.63 µs 196.21 µs 196.80 µs]
                        change: [+238.34% +240.36% +242.05%] (p = 0.00 < 0.05)
                        Performance has regressed.

     Running benches/nybb.rs (target/release/deps/nybb-2866c0cc69c207fd)
Gnuplot not found, using plotters backend
to geo                  time:   [968.54 µs 976.46 µs 985.41 µs]
                        change: [+372.81% +376.08% +380.35%] (p = 0.00 < 0.05)
                        Performance has regressed.

Benchmarking euclidean distance to scalar point: Warming up for 3.0000 s
Warning: Unable to complete 100 samples in 5.0s. You may wish to increase target time to 6.8s, enable flat sampling, or reduce sample count to 60.
euclidean distance to scalar point
                        time:   [1.3467 ms 1.3489 ms 1.3511 ms]
                        change: [+120.24% +120.80% +121.32%] (p = 0.00 < 0.05)
                        Performance has regressed.

     Running benches/translate.rs (target/release/deps/translate-4c10f070b3d44d11)
Gnuplot not found, using plotters backend
translate PolygonArray  time:   [241.38 µs 242.12 µs 243.02 µs]
                        change: [+140.10% +140.91% +142.07%] (p = 0.00 < 0.05)
                        Performance has regressed.