Attempt structuredClone in cloneDeep before doing manual clone

[TheJaredWilcurt]

I would assume that the structuredClone API being native would allow JS engine makers to optimize the performance. Though it isn't as powerful as Lodash's clone deep, that power may not be needed in all use cases.

My suggestion is basically this:

function cloneDeep (value) {
  try {
    if (structuredClone) {
      return structuredClone(value);
    }
  } catch {
    return doWhatLodashHasAlwaysDoneForCloneDeep(value);
  }
}

Just an idea, not sure if this would be a net positive or negative (more people dealing with the time it takes to do the failed attempt before falling back, versus more people getting the fast native solution and skipping the more complex JS solution).

Alternatively every user could just make their own helper function like this:

import { cloneDeep } from 'lodash.cloneDeep';

export const deepClone = function (value) {
  try {
    if (structuredClone) {
      return structuredClone(value);
    }
  } catch {
    return cloneDeep(value);
  }
};

But the whole point of Lodash it to not need to make/copy a bunch of common helpers for stuff like this.

[jdalton]

YES YES YES! @TheJaredWilcurt up for creating a PR?

[kapilkumar9395]

How we are going to check for complex object? Structured Clone Algorithm doesn't support functions and other things.

[Trott]

How we are going to check for complex object? Structured Clone Algorithm doesn't support functions and other things.

It seems like this question is answered in the code suggestion by OP. Trying to clone a function will throw, and fall back to the current implementation.

[kapilkumar9395]

That is correct. I overlooked try catch.

[TheJaredWilcurt]

YES YES YES! @TheJaredWilcurt up for creating a PR?
- @jdalton

Unfortunately I don't think I will be able to any time soon. Despite what the VC-Backed Bun claims, it in fact does not actually run on Windows. I don't know if I'll have time to clone lodash down on one of my Linux VMs and also try to get bun working on them.

---

Implementation Ideas

I would request, for whoever does implement this feature, that you allow passing in an option into cloneDeep to skip the native check. Since sometimes I have objects I know will fail the native try/catch, so I could save some time by skipping that attempt. I imagine others would appreciate this option too.

[lewxdev]

YES YES YES! @TheJaredWilcurt up for creating a PR?

• @jdalton

Unfortunately I don't think I will be able to any time soon. Despite what the VC-Backed Bun claims, it in fact does not actually run on Windows. I don't know if I'll have time to clone lodash down on one of my Linux VMs and also try to get bun working on them.

Implementation Ideas

I would request, for whoever does implement this feature, that you allow passing in an option into cloneDeep to skip the native check. Since sometimes I have objects I know will fail the native try/catch, so I could save some time by skipping that attempt. I imagine others would appreciate this option too.

I can try implementing this PR 👍🏽