Dropping into self

Describe the bug You can drag an item into it self if it has a child that accepts drop.

https://p195.p4.n0.cdn.getcloudapp.com/items/geud4ggR/d26a0b2d-f670-4252-8347-ef92a302968b.gif?v=907ccf16be512a91e317756cf823cf47

Expected Disallow this sort of thing.

canDropAt={(items, target) => {

// prevent self drop
if(items[0].index == target.parentItem) return false

            // check if target is a child of parent
            const isSelf = itemsSource[items[0].index].children.findIndex(i => i == target.parentItem)
            if(isSelf !== -1) return false
          
return true
 } 
}

That is some basic code to check. If the fist item is the same as the a parent. I also check if parentItem is a child of the first item checking the items. (I called it itemsSource).

This isn’t great and doesn’t check for multiple selects. But it kind of works. 😃

[lukasbach]

Hi @C7X! When working with lists with thousands of entries, list virtualization starts to become inevitable. This is something I want to support with react-complex-tree in the future, but there is still a bigger road ahead for that and that will not come with the performance fixes I am working on right now, sorry. So for now, I unfortunately have to say that other libraries might be a better starting point until rct supports virtualization.

[lukasbach]

Hi, thanks for reporting! Just as an heads up, I’ve looked into it, and it is not quite as trivial to fix; since checking whether someone is dropping an item into itself is not a trivial operation (can be a high depth through which is being dropped, and arbitrarily many items can be dragged at once, so the check is required for every dragged item), this will require some calculation. And, currently, the drop check happens on the drag event, not the drop event (so that the library can already prevent dropping on not allowed items), so this might not scale well for bigger trees. I’m currently working on improving the overall performance of the library, and then plan to only do the check during the initial drag event, hopefully in a way that only requires linear effort. Until then, a workaround similar to what @impactvelocity suggested is probably the easiest way to go. I’ll keep you updated on the progress!