lib: replace public Map methods with primordials

[aduh95]

Checklist

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• commit message follows commit guidelines

[RaisinTen]

Does this need to be benchmarked against process?

[aduh95]

Benchmark CI: https://ci.nodejs.org/view/Node.js%20benchmark/job/benchmark-node-micro-benchmarks/895/

EDIT: Didn't show any perf regression or improvement 👍

Benchmark results

                                                      confidence improvement accuracy (*)    (**)   (***)
 process/bench-env.js operation='delete' n=1000000                   -0.19 %       ±4.42%  ±5.91%  ±7.75%
 process/bench-env.js operation='enumerate' n=1000000                 0.01 %       ±1.45%  ±1.93%  ±2.51%
 process/bench-env.js operation='get' n=1000000                      -2.35 %       ±5.17%  ±6.89%  ±8.99%
 process/bench-env.js operation='query' n=1000000                    -2.14 %       ±9.47% ±12.61% ±16.44%
 process/bench-env.js operation='set' n=1000000                      -1.77 %       ±6.70%  ±8.91% ±11.60%
 process/bench-hrtime.js type='bigint' n=1000000                      3.42 %       ±5.66%  ±7.53%  ±9.81%
 process/bench-hrtime.js type='diff' n=1000000                        2.72 %       ±3.48%  ±4.62%  ±6.02%
 process/bench-hrtime.js type='raw' n=1000000                         0.46 %       ±4.47%  ±5.95%  ±7.76%
 process/memoryUsage.js n=100000                                     -0.53 %       ±3.34%  ±4.46%  ±5.84%
 process/next-tick-breadth-args.js n=10000000                         2.98 %       ±4.18%  ±5.56%  ±7.24%
 process/next-tick-breadth.js n=10000000                              0.02 %       ±4.37%  ±5.82%  ±7.60%
 process/next-tick-depth-args.js n=7000000                            1.40 %       ±4.55%  ±6.07%  ±7.92%
 process/next-tick-depth.js n=7000000                                 0.88 %       ±3.53%  ±4.71%  ±6.18%
 process/next-tick-exec-args.js n=4000000                            -4.74 %      ±11.85% ±15.79% ±20.59%
 process/next-tick-exec.js n=4000000                                  4.45 %       ±6.18%  ±8.23% ±10.71%
 process/queue-microtask-breadth.js n=400000                         -4.26 %       ±7.29%  ±9.70% ±12.63%
 process/queue-microtask-depth.js n=1200000                          -1.21 %       ±1.53%  ±2.04%  ±2.66%
 process/resourceUsage.js n=100000                                    1.65 %       ±3.25%  ±4.33%  ±5.66%

Be aware that when doing many comparisons the risk of a false-positive
result increases. In this case there are 18 comparisons, you can thus
expect the following amount of false-positive results:
  0.90 false positives, when considering a   5% risk acceptance (*, **, ***),
  0.18 false positives, when considering a   1% risk acceptance (**, ***),
  0.02 false positives, when considering a 0.1% risk acceptance (***)

[ExE-Boss]

Alternatively, it might be better to fix the C++ binding, so that getOptions() returns a SafeMap, just like how the C++ API handles creation of Buffer instances.

[aduh95]

Alternatively, it might be better to fix the C++ binding, so that getOptions() returns a SafeMap

Agreed, having to deal only with SafeMap objects rather than Map would be ideal. That being said, getOptionValue returns simple string, so most core developers shouldn't have to deal with the options object directly.

just like how the C++ API handles creation of Buffer instances.

I've been looking into that, and it seems all the C++ API does is create a Uint8Array and change the object's prototype:

node/src/node_buffer.cc

Lines 275 to 277 in 7efada6

	Local<Uint8Array> ui = Uint8Array::New(ab, byte_offset, length);
	Maybe<bool> mb =
	ui->SetPrototype(env->context(), env->buffer_prototype_object());

This could be done in JS, no need to change the C++, but I wouldn't be surprised if it had worse performance (haven't measured it though).

[aduh95]

Superseded by #36989 and #37029.