TreeGrid loads all data from the data provider when scroll position changes

For fetching hierarchical data (for TreeGrid) the HierarchyMapper is used. Method

public Stream<T> fetchItems(Range range)

is invoked every time the current scroll position of the Tree(Grid) changes. What happens is that the hierarchy is retrieved, and then in memory stream operations are done to skip to and limit the correct amount of data shown like this:

return getHierarchy(null).skip(range.getStart()).limit(range.length());

This will effectively for every scroll position change load ALL data from the backing data provider. Needless to say this will slow down any lazy loading intended hierarchical grids to crawl if there's more than trivial amount of data.

Start with HierarchyMapper.fetchItems(Range range) and trace its getHierarchy(null) call. Reproduce scrolling with a lazy-loading hierarchical data provider and inspect whether the full hierarchy is loaded before skip and limit. Done means range fetching no longer loads all backing data on each scroll position change.