perf(core): convert `for..in` to `Object.keys().forEach` for better perf

[ghiscoding]

• as mentioned in this Stack Overflow, Object.keys().forEach are quicker than for..in because the latter also loops through prototype chain.