fs: optimize fs.cpSync js calls

[anonrig]

#53612 is required to do a benchmark

A different approach to #53541

cc @nodejs/performance @nodejs/cpp-reviewers @billywhizz @jasnell @nodejs/fs

Benchmark CI: https://ci.nodejs.org/view/Node.js%20benchmark/job/benchmark-node-micro-benchmarks/1573/

Small win on happy path, and probably high win on error path:

                           confidence improvement accuracy (*)   (**)   (***)
fs/bench-cpSync.js n=1            ***     21.96 %       ±6.85% ±9.14% ±11.97%
fs/bench-cpSync.js n=100          ***     14.91 %       ±2.86% ±3.81%  ±4.97%
fs/bench-cpSync.js n=10000        ***      9.74 %       ±1.28% ±1.71%  ±2.23%

[anonrig]

cc @nodejs/tsc @nodejs/fs appreciate your input on which path to take.

previous pr:

• is a sem-ver major
• introduces 2 different implementations (js only and cpp)
• had limitations due to std::filesystem::copy

this pr:

• is not a sem-ver major
• optimizes all flags passed to cpSync
• optimizes error cases as same as the previous pr
• has a lot more optimization to do

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/60000/

[mcollina]

lgtm