feat(util): use native Map if available

[JonasBa]

Related to the discussion in #965

This PR is lacking tests, if the general direction is ok and validated, I will add the relevant tests.

From initial benchmarks

> node src/core/util.benchmark.js
cpu: Apple M1 Pro
runtime: node v18.3.0 (arm64-darwin)

benchmark      time (avg)             (min … max)       p75       p99      p995
------------------------------------------------- -----------------------------
• n=8
------------------------------------------------- -----------------------------
Native      92.59 ns/iter  (88.37 ns … 521.94 ns)  94.09 ns  116.1 ns 131.65 ns
Polyfill   140.06 ns/iter (137.22 ns … 171.45 ns) 139.03 ns 164.76 ns 170.73 ns

summary for n=8
  Native
   1.51x faster than Polyfill

• n=32
------------------------------------------------- -----------------------------
Native      298.8 ns/iter    (292 ns … 649.44 ns) 294.09 ns 630.26 ns 649.44 ns
Polyfill     1.17 µs/iter     (1.15 µs … 1.19 µs)   1.17 µs   1.19 µs   1.19 µs

summary for n=32
  Native
   3.9x faster than Polyfill

• n=64
------------------------------------------------- -----------------------------
Native     558.55 ns/iter (554.65 ns … 581.92 ns)  558.3 ns 580.21 ns 581.92 ns
Polyfill      2.4 µs/iter     (2.38 µs … 2.43 µs)   2.41 µs   2.43 µs   2.43 µs

summary for n=64
  Native
   4.3x faster than Polyfill

• n=128
------------------------------------------------- -----------------------------
Native       1.09 µs/iter     (1.08 µs … 1.13 µs)   1.09 µs   1.13 µs   1.13 µs
Polyfill      4.8 µs/iter     (4.76 µs … 4.84 µs)   4.81 µs   4.84 µs   4.84 µs

summary for n=128
  Native
   4.41x faster than Polyfill

• n=1024
------------------------------------------------- -----------------------------
Native       8.47 µs/iter     (8.46 µs … 8.57 µs)   8.47 µs   8.57 µs   8.57 µs
Polyfill    43.66 µs/iter  (42.83 µs … 236.92 µs)  43.54 µs  46.92 µs  48.67 µs

summary for n=1024
  Native
   5.15x faster than Polyfill

• n=8192
------------------------------------------------- -----------------------------
Native      68.23 µs/iter    (67.5 µs … 93.42 µs)  68.25 µs  72.46 µs  73.46 µs
Polyfill   771.05 µs/iter   (752.29 µs … 1.06 ms) 772.42 µs 871.79 µs 897.46 µs

summary for n=8192
  Native
   11.3x faster than Polyfill

• n=200000
------------------------------------------------- -----------------------------
Native       1.66 ms/iter     (1.64 ms … 1.71 ms)   1.66 ms    1.7 ms    1.7 ms
Polyfill    28.46 ms/iter    (28.15 ms … 29.6 ms)  28.51 ms   29.6 ms   29.6 ms

summary for n=200000
  Native
   17.17x faster than Polyfill

[plainheart]

The others LGTM

[plainheart]

Thank you @JonasBa for the nice work!
@pissang Any other thoughts? If none, we can merge it.

[JonasBa]

Appreciate the thorough review and helpful comments @plainheart @pissang 🙏