Optimize Kotlin primary constructor fixture generation

[char-yb]

Summary

This PR optimizes Kotlin primary constructor-based fixture generation without adding new dependencies or changing public APIs.

It reduces per-sample overhead in Kotlin object creation paths, which can become more visible when tests run under JVM instrumentation tools such as Kover.

(Optional): Description

The previous implementation always used callBy(...) and built intermediate maps for every generated Kotlin object. callBy(...) is still necessary when optional constructor parameters should be omitted, but it is more expensive than needed when all parameters are already available.

This PR keeps the existing optional-parameter behavior intact and only uses the faster KFunction.call(*args) path when every constructor parameter can be provided.

Changes include:

• Reduce repeated actualType() / cachedKotlin() calls in PrimaryConstructorArbitraryIntrospector.match().
• Avoid mapKeys allocation in the constructor invocation path by building the name-value map directly.
• Use KFunction.call(*args) when no optional constructor parameter needs to be skipped.
• Keep callBy(...) when optional parameters must be skipped.
• Preserve explicit null behavior for nullable optional parameters.
• Change constructor parameter name lookup from List.contains to Set.contains.
• Add regression tests for optional default values and nullable optional default values.

Local quick JMH benchmark:

Version	Score
Before	1214.915 ms/op
After	904.979 ms/op

Benchmark command:

./gradlew :fixture-monkey-benchmarks:fixture-monkey-benchmark-kotlin:jmh \
  -Pjmh.quick \
  '-Pjmh.include=.*primaryConstructorGenerateKotlinOrderSheetWithFixtureMonkey.*'

How Has This Been Tested?

./gradlew :fixture-monkey-kotlin:test :fixture-monkey-kotlin:ktlintCheck
./gradlew :fixture-monkey-tests:kotlin-tests:test

I also verified the benchmark above locally to compare the Kotlin primary constructor fixture generation path before and after this change.

Is the Document updated?

No.

This PR does not change public APIs, user-facing behavior, configuration, or documented usage. It only optimizes the existing Kotlin primary constructor generation path.

Please feel free to share any feedback, especially if there are concerns about the constructor invocation behavior or benchmark scope.

[CLAassistant]

All committers have signed the CLA.

[seongahjo]

👍