@maaslalani @meowgorithm (apologies for pinging you directly, but I saw you are members of Charm that are active in the issues section).

Is there an interest in my proposal, or should I just keep it as an out of tree model?

@mariusor Apologies for not having responded sooner, but we'll definitely check out your code! A tree-view bubble is probably inevitable and we definitely see a need for such a component. It's certainly one of the most complex visual components and getting it right is non-trivial, so I'm excited to see what you came up with!

Thank you @muesli. Looking forward to comments.

Hello everyone, any hope to review my tree model?

I have made some improvements that extend the use to something like a threaded "conversation" model, which has another example

Look awesome!

Hey, thanks for this! I saw your messages on Mastodon as well. We're definitely interested in your implementation, we just need to give it a very thorough review and try to break it. It may take a while to merge, but please don't take that as a lack of interest. Thank you so much for your patience!

Hi, I think that with the addition of the lipgloss/tree model my PR is probably unnecessary to merge.

I have not played with that one to see if there are any performance or any significant API differences but I trust you guys.

Hey again @mariusor: just a heads up that @dlvdhr is tackling this one in #639, which we intend to merge.

No worries. I have resigned myself that my model won't make it into the official repository.

I like my implementation better with regards to having the Node type be an interface that any other model can implement and be part of a tree, but I'm not married to the idea, it's probably much simpler having it as a stand alone data struct.

Having it as an interface allows that the logic of something like expand/collapse, multi-line, different sizes/heights for the nodes, etc. can be part of the individual node models instead of requiring the tree to have code for each custom behaviour. I haven't hit any performance issues yet due to this (propagating state to the nodes needs to be done with Updates for all nodes), but I'm not rendering very deep/tall trees.

I'm not sure I'm going to change my own applications that make use of trees to this new implementation, but that's mostly due to laziness. :)