Avoid dynamic dispatch when calling wrapCheck

[yepitschunked]

Q	A
Fixed Issues?	N/A
Patch: Bug Fix?	Yes
Major: Breaking Change?
Minor: New Feature?
Tests Added + Pass?	Yes
Documentation PR Link
Any Dependency Changes?
License	MIT

Micro-optimization to avoid dynamic function dispatch in wrapCheck, which is in the hot path of babel transforms. The current code interpolates a string and looks up a function at runtime during every visitor function invocation, whereas this PR looks up the function once while building the visitor and just calls it during visitor function invocation.

I couldn't find any docs on babel benchmarking, so I used swc's benchmark and modified it to only run babel. Here are the before/after results:

Before:

[es3] Running single thread benchmarks
babel (preset-env) x 149 ops/sec ±2.92% (79 runs sampled)
[es3] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es5] Running single thread benchmarks
babel (preset-env) x 159 ops/sec ±2.71% (83 runs sampled)
[es5] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es2015] Running single thread benchmarks
babel (preset-env) x 162 ops/sec ±2.29% (76 runs sampled)
[es2015] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es2016] Running single thread benchmarks
babel (preset-env) x 163 ops/sec ±2.24% (77 runs sampled)
[es2016] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es2017] Running single thread benchmarks
babel (preset-env) x 162 ops/sec ±2.77% (84 runs sampled)
[es2017] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es2018] Running single thread benchmarks
babel (preset-env) x 162 ops/sec ±2.78% (84 runs sampled)
[es2018] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es2019] Running single thread benchmarks
babel (preset-env) x 161 ops/sec ±2.56% (84 runs sampled)
[es2019] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es2020] Running single thread benchmarks
babel (preset-env) x 163 ops/sec ±2.25% (84 runs sampled)
[es2020] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)

After:

[es3] Running single thread benchmarks
babel (preset-env) x 160 ops/sec ±3.17% (83 runs sampled)
[es3] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es5] Running single thread benchmarks
babel (preset-env) x 169 ops/sec ±2.78% (79 runs sampled)
[es5] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es2015] Running single thread benchmarks
babel (preset-env) x 171 ops/sec ±2.28% (81 runs sampled)
[es2015] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es2016] Running single thread benchmarks
babel (preset-env) x 170 ops/sec ±2.96% (81 runs sampled)
[es2016] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es2017] Running single thread benchmarks
babel (preset-env) x 171 ops/sec ±2.88% (81 runs sampled)
[es2017] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es2018] Running single thread benchmarks
babel (preset-env) x 173 ops/sec ±2.61% (82 runs sampled)
[es2018] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es2019] Running single thread benchmarks
babel (preset-env) x 173 ops/sec ±2.37% (81 runs sampled)
[es2019] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)
[es2020] Running single thread benchmarks
babel (preset-env) x 171 ops/sec ±2.53% (81 runs sampled)
[es2020] Transform rxjs/Observable.ts benchmark bench suite: Fastest is babel (preset-env)

It's roughly a 5-6% improvement.

[JLHwung]

Thanks!

[nicolo-ribaudo]

Thank you!