Improve performance of removing elements from sequences

[jeff303]

Merging InsertBeforeIndex protocol into new IndexedOps protocol for optimized indexed based operations (which now includes inserting before and removing at)

Updating nthpath and keypath to use new helper fns

Adding tests and benchmarks

[jeff303]

I think this change should also cover must because it calls do-keypath-transform. Output of new benchmarks from my machine:

Benchmark: set keypath and nthpath at index to NONE versus srange in middle (vector)

Mean(us)	vs best		Code
31.2 		 1.00 		 (setval (nthpath middle-idx) NONE data-vec)
32.1 		 1.03 		 (setval (keypath middle-idx) NONE data-vec)
145 		 4.65 		 (setval (srange middle-idx (inc middle-idx)) [] data-vec)

********************************

Benchmark: set keypath and nthpath at index to NONE versus srange in middle (list)

Mean(us)	vs best		Code
5.56 		 1.00 		 (setval (keypath middle-idx) NONE data-lst)
12.5 		 2.24 		 (setval (srange middle-idx (inc middle-idx)) [] data-lst)
12.7 		 2.29 		 (setval (nthpath middle-idx) NONE data-lst)

********************************

Benchmark: set keypath and nthpath at beginning to NONE versus srange and subvec (vector)

Mean(us)	vs best		Code
0.0179 		 1.00 		 (subvec data-vec 1)
0.150 		 8.39 		 (setval (keypath 0) NONE data-vec)
0.160 		 8.89 		 (setval (nthpath 0) NONE data-vec)
107 		 5.98e+03 		 (setval (srange 0 1) [] data-vec)

********************************

Benchmark: set keypath and nthpath at beginning to NONE versus srange and rest (list)

Mean(us)	vs best		Code
0.00733 		 1.00 		 (rest data-lst)
0.192 		 26.3 		 (setval (keypath 0) NONE data-lst)
12.5 		 1.71e+03 		 (setval (srange 0 1) [] data-lst)
12.8 		 1.75e+03 		 (setval (nthpath 0) NONE data-lst)

********************************

Benchmark: set keypath and nthpath at end to NONE versus srange and subvec (vector)

Mean(us)	vs best		Code
0.0148 		 1.00 		 (subvec data-vec 0 (dec size))
0.257 		 17.4 		 (setval (keypath (dec size)) NONE data-vec)
0.275 		 18.6 		 (setval (nthpath (dec size)) NONE data-vec)
194 		 1.31e+04 		 (setval (srange (dec size) size) [] data-vec)

********************************

[jeff303]

This should cover most of #219, although it doesn't do anything to address srange performance.