Revert last changes that caused performance regressions

[stephenamar-db]

I don't know why the Make evalutator a simple val made a difference, but it did.

We are back to the performance level of the scala3 migration (which is better than the 0.4 series)

# Run progress: 0.00% complete, ETA 00:02:00
# Fork: 1 of 1
Java HotSpot(TM) 64-Bit Server VM warning: -XX:ThreadPriorityPolicy=1 may require system level permission, e.g., being the root user. If the necessary permission is not possessed, changes to priority will be silently ignored.
# Warmup Iteration   1: 318.899 ms/op
# Warmup Iteration   2: 219.683 ms/op
Iteration   1: 218.122 ms/op
Iteration   2: 217.369 ms/op
Iteration   3: 216.576 ms/op
Iteration   4: 216.465 ms/op
Iteration   5: 215.468 ms/op
Iteration   6: 215.459 ms/op
Iteration   7: 215.341 ms/op
Iteration   8: 214.828 ms/op
Iteration   9: 215.898 ms/op
Iteration  10: 215.108 ms/op


Result "com.databricks.jsonnet.bundle.test.InternalJsonnetBuilderBenchmark.main":
  216.063 ±(99.9%) 1.600 ms/op [Average]
  (min, avg, max) = (214.828, 216.063, 218.122), stdev = 1.058
  CI (99.9%): [214.464, 217.663] (assumes normal distribution)


# Run complete. Total time: 00:02:02

REMEMBER: The numbers below are just data. To gain reusable insights, you need to follow up on
why the numbers are the way they are. Use profilers (see -prof, -lprof), design factorial
experiments, perform baseline and negative tests that provide experimental control, make sure
the benchmarking environment is safe on JVM/OS/HW level, ask for reviews from the domain experts.
Do not assume the numbers tell you what you want them to tell.

Benchmark                             Mode  Cnt    Score   Error  Units
InternalJsonnetBuilderBenchmark.main  avgt   10  216.063 ± 1.600  ms/op

[stephenamar-db]

@He-Pin

[He-Pin]

Wired, The lazy val will access the barrier

[He-Pin]

If that's the lazy val bootst the performance, please add an comments.

[He-Pin]

If this could work, then a var evaluater and evaluater() method to get or create will works too , and avoid the double check, wired any way.