Extent is broken?

[nemosmithasf]

If a tree has the extent [[1,1],[5,5]], calling tree.extent( tree.extent() ) returns a tree with the extent [[1,1], [9,9]]

[Fil]

A point that is on the upper limit (of x or of y) belongs to the quad that is above it.

d3.quadtree([[0, 0]]).cover(0.99999999, 0).extent() // [[0,0],[1,1]]  
d3.quadtree([[0, 0]]).cover(1, 0).extent() // [[0,0],[2,2]]

but now that you're pointing at this, is this sentence correct?
“x0 and y0 are the inclusive lower bounds and x1 and y1 are the inclusive upper bounds”
https://github.com/d3/d3-quadtree#quadtree_extent ; shouldn't it be "exclusive" for the upper bounds?

[nemosmithasf]

That's what I assumed, but isn't it count-intuitive for tree.extent( tree.extent() ) to return a different result?

The main problem is that it makes serializing/unserializing the quadtree very awkward, because there's no other way to set the extent that doesn't reach into private properties (e.g. _x0, _y0)