ARROW-11107: [Rust] Faster Length kernel (10%)

[jorgecarleitao]

This is a corollary of the API PRed in #9076 , and how we can use it to efficiently build Buffer without even using the MutableBuffer. The main idea here is that we can write

 let lengths: Buffer = slice
.windows(2)
.map(|offset| offset[1] - offset[0])
.collect();

and this (efficiently) builds a Buffer from the offset differences.

Bench:

length time: [3.2345 us 3.2457 us 3.2602 us] change: [-12.072% -11.361% -10.654%] (p = 0.00 < 0.05)
Performance has improved.
Found 10 outliers among 100 measurements (10.00%)
4 (4.00%) high mild
6 (6.00%) high severe

[github-actions]

https://issues.apache.org/jira/browse/ARROW-11107

[codecov-io]

Codecov Report

Merging #9077 (c1badea) into master (4b7cdcb) will decrease coverage by 0.05%.
The diff coverage is 93.50%.

@@ Coverage Diff @@## master #9077 +/- ##
==========================================
- Coverage 82.61% 82.56% -0.06% 
==========================================
Files 203 204 +1 Lines 50140 50091 -49 ==========================================
- Hits 41422 41356 -66 - Misses 8718 8735 +17 

Impacted Files	Coverage Δ
rust/arrow/src/bytes.rs	53.12% <ø> (-5.21%)	⬇️
rust/arrow/src/compute/kernels/aggregate.rs	74.93% <ø> (-0.07%)	⬇️
rust/arrow/src/datatypes.rs	75.02% <ø> (ø)
rust/arrow/src/array/transform/list.rs	86.20% <66.66%> (+2.33%)	⬆️
rust/arrow/src/array/transform/mod.rs	88.29% <83.33%> (ø)
rust/arrow/src/array/builder.rs	85.23% <86.79%> (-0.67%)	⬇️
rust/arrow/src/buffer.rs	95.53% <88.75%> (-2.19%)	⬇️
rust/arrow/src/array/array_binary.rs	90.54% <100.00%> (ø)
rust/arrow/src/array/array_list.rs	92.66% <100.00%> (-0.45%)	⬇️
rust/arrow/src/array/array_primitive.rs	92.28% <100.00%> (-0.04%)	⬇️
... and 25 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4b7cdcb...c1badea. Read the comment docs.