perf: intern identifier field names for faster field lookup

[He-Pin]

Motivation

Object field lookup is the hottest path in sjsonnet evaluation. Every field access (containsKey, containsVisibleKey, valueRaw) compares String keys char-by-char via .equals, which is wasteful when the same field names repeat across objects (common in K8s manifests, stdlib).

Modification

• Parser routes identifier field names (fieldname, Select) through internedStrings, sharing String instances across repeated parses (same > 1024 length guard as the existing string-literal interning).
• With interned keys, String.equals hits its built-in reference check on the first line, so lookups short-circuit without char-by-char comparison.

An earlier revision also prefixed the lookup loops with an explicit (x eq k) || fast path; it was dropped because String.equals already begins with an identity check on JVM and Scala Native, and on Scala.js both eq and equals compile to ===, so the prefix double-compared every non-matching key.

Result

Benchmark	master	this PR	Delta
ParserBenchmark.main	1.462 ± 0.053 ms/op	1.378 ± 0.014 ms/op	-5.7%
MainBenchmark.main	3.028 ± 0.536 ms/op	3.085 ± 0.529 ms/op	+1.9% (noise)

JMH config: -f 2 -wi 5 -i 10 -w 1 -r 1 (2 forks × 10 measurement iterations)

All tests pass. Zero behavioral change — interning only affects String identity, never equality semantics.

Test plan

• ./mill 'sjsonnet.jvm[_]'.test — all pass
• ./mill bench.runJmh ".*MainBenchmark.*" — no regression
• ./mill bench.runJmh ".*ParserBenchmark.*" — 5.7% improvement