Optimize parsing performance

[intsuc]

Overview

This optimization avoids unnecessary object allocations by not taking snapshots of the parsing state when the size of the relevant nodes is 1. Parsing time becomes about 48.6 to 57.1% of the original.

Benchmark Results

Before Optimization f20bede

Benchmark	Mode	Cnt	Score	Error	Units
ParsingBenchmarks.parse_	avgt	25	102.391	± 1.121	ns/op
ParsingBenchmarks.parse_a1i	avgt	25	322.832	± 3.665	ns/op
ParsingBenchmarks.parse_c	avgt	25	102.086	± 1.063	ns/op
ParsingBenchmarks.parse_k1i	avgt	25	338.494	± 2.231	ns/op
ParsingBenchmarks.parse_l123456789	avgt	25	1014.364	± 9.041	ns/op

After Optimization f4d27ab

Benchmark	Mode	Cnt	Score	Error	Units
ParsingBenchmarks.parse_	avgt	25	52.451	± 0.291	ns/op
ParsingBenchmarks.parse_a1i	avgt	25	164.723	± 1.224	ns/op
ParsingBenchmarks.parse_c	avgt	25	52.949	± 0.291	ns/op
ParsingBenchmarks.parse_k1i	avgt	25	193.371	± 1.875	ns/op
ParsingBenchmarks.parse_l123456789	avgt	25	492.890	± 5.163	ns/op

[intsuc]

If unique, we do not need to take snapshots of the parsing state (contextSoFar and originalReader) here.

[intsuc]

If unique, the potentials will always contain only parse and this parse will be returned after exiting this for loop, so we can return parse here without allocating potentials.

[intsuc]

Since potentials will contain ParseResults<S> up to the amount of relevantNodes.size(), allocate that capacity in advance to ensure resizing never occurs.

[yanisderbikov]

I think this is less readable. But if it increase performance it's ok.

[yanisderbikov]

I think this is less readable. But if it increase performance it's ok.