Adapt zFCP UI to HTTP API v2

[joseivanlopez]

Revamp the zFCP UI: adapt to HTTP API v2 and follow the same patterns as rest of the UI.

[dgdavid]

Thanks a ton for taking care of this 🙏 I know it has been a little headache 😓, but the result will pay-off. I have to check to minor things before merging it, reason why I'm postponing the approval.

[jreidinger]

well, I would probably use StructuredClone as it is already enough supported - https://caniuse.com/?search=structuredClone

[dgdavid]

As said in IRC, let's postpone such a big decision until we have more time to play and learn about such a method. I prefer to be a bit conservative here 🙏

From MDN.

The method also allows transferable objects in the original value to be transferred rather than cloned to the new object. Transferred objects are detached from the original object and attached to the new object; they are no longer accessible in the original object.

https://developer.mozilla.org/en-US/docs/Web/API/Window/structuredClone

[jreidinger]

also what about nullish operator? So something like
return structuredClone(object ?? DEFAULT_CONFIG)

[jreidinger]

here is link why destructuring is not enough if object contain array - https://runjs.app/play/#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

[dgdavid]

I see your point. However, I think in our case this shouldn't be a concern because we're not destructuring objects in a way that would lead to this situation. The pattern we're using is slightly different: with each new copy we're "replacing the previous state" rather than mutating or reusing it. Not sure if I'm explaining it well.

So while it's true that destructuring only creates a shallow copy (and arrays or nested objects will still share references), that scenario shouldn't occur with how we're handling the data here.

That said, it's always good to stay vigilant about these edge cases and improve things, so thanks for pointing it out 👍

[dgdavid]

I've sent aesthetic changes, find screenshots attached below. Let me know if you find them to divergent with your view. If not, you can go ahead with the merge. I think the recursive fix can be postponed for another round of improvements, but up to you to go for the reducer proposed by @jreidinger before merging. No strong opinion