Scroller: dynamic size for itemSize #2290

[Max0nyshchenko]

This is the example PR to the discussion in https://github.com/orgs/primefaces/discussions/2290

These changes need to be further optimized, which I can do, if you are going to accepts the changes, but they already working and you can see and check them out in the docs examples that has been added (flexible-items-height-doc and flexible-grid-doc). Or you can finish this PR by yourself, feel free to use my changes in any way you want. The main thing is to have the support for flexible itemSize for different items.

I need these changes so that I can implement expandable rows in p-table with virtualization later, which is required for one of my projects.
So I'm wondering if you consider these changes significant or of the kind that would be welcomed.
If you give me the green light I'll work on the final version with improved performance.

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

2 Skipped Deployments

Name	Status	Preview	Updated (UTC)
primeng	⬜️ Ignored (Inspect)	Visit Preview	Dec 25, 2024 8:10pm
primeng-v18	⬜️ Ignored (Inspect)	Visit Preview	Dec 25, 2024 8:10pm

[DanniKach]

Oh, thanks a lot @Max0nyshchenko.
We really researched a lot of topics before understood that such functionality is unavailable in current version of library, so it is impossible to implement expanded items with infinite scroll. This is required functionality for our application, so hope it will be approved and merged as soon as possible. We are waiting for it.

[Max0nyshchenko]

@cetincakiroglu what do you think? Could you take a look?

[mertsincan]

Hi @Max0nyshchenko,
Thanks a lot for your contribution! This seems like a draft. Are you still working on it?

[Max0nyshchenko]

Hi @mertsincan ,
You're welcome. It is a draft cause I need to know if this kind of changes will be accepted so I can see if it makes sense for me to finish it.
The only thing left to finish is performance optimization. Currently, the recalculateSize function is called on every scroll event, it is computationally-intensive function and can take some time to execute in large data sets.
I'm thinking of introducing caching of recalculateSize results and scroll event debouncing (or maybe using requestAnimationFrame instead of scroll event) for optimization purposes. But maybe you have better suggestions?

So 2 questions:
1 - Will these changes be welcomed (maybe you don't approve of an interface or don't want to have such a feature at all, as by your policies you don't accept features)?
2 - Any ideas for optimization?

[vercel]

@Max0nyshchenko is attempting to deploy a commit to the primetek Team on Vercel.

A member of the Team first needs to authorize it.