While working on the tree control for our extension, I noticed that the sample implementation was not optimal. For instance, `vscode.TreeDataProvider.onDidChangeTreeData` should only be called once for the top-most node that has changed, as all children will be refreshed as well.
This change updates the code of the sample to track state changes during propagation and only raise the event for the top-most parent node that has its state changing.
There also was an issue with root node `1` not rendering, which impacted the new algorithm, as it was not refreshing the tree, when invisible root was changing. In order to address this, I updated `getChildren` to actually return the `1` node, when root was requested. This in turn changed the presentation (`1` is now rendered in the tree) and addressed the problem with tree not refreshing when `1` is the node that raises the event.
Lastly, there is one more optimization that I didn't make due to the bug that is in the works: `getTreeItem` can actually return the `element` directly, without the copy, as it already implements `TreeComponentItem`. The problem currently is that `TreeNode` also defines an `id`, which causes an error during refresh. Once that is addressed - this part can be optimized as well.
- Upate vscode-nls to 4.0.0 in notebook extension. Should be fix for Insiders build failure due to localization package failing
- Updated samples to remove as vscode-nls if not needed, or update if still needed
- Updated samples using `npm audit fix`
- Fixed compile errors in all the samples
