That's a fair question. Hard to say, but my guess is that if we do the
first, the incremental improvements of the second might not be worth the
effort or risk.
And if we go with the second, we would need another test case that hits
the slow iteration problem when removing a listener from a
Assuming testing and performance/memory analysis leads to the conclusion that
the risks are worth it, would it make sense to do both? Would we get a greater
benefit from the combined effects? Or is the incremental improvement of
including the second fix (whichever it may be) no longer
We now have two pull requests under review that propose to solve the
poor scrolling performance of TableView and TreeTableView, as tracked by
JDK-8185886 [1]
The first one, PR #108 [2], proposes a change in the bindings
ExpressionHelper code relating to the cleaning up of listeners (changing
