On Thu, 10 Nov 2022 16:44:06 GMT, Marius Hanl <mh...@openjdk.org> wrote:
>> I looked at the other layout panes and they already do something similar to >> what this PR does, using the snapped insets for intermediate compuation. I >> also note that some of the computation is calling superclass methods that do >> snapping already, and expect it to be added to snapped insets (if there are >> insets), for example, `computeWidth()`. In this particular case -- >> `snappedLeftInset() + leftAnchor` -- I don't think `leftAnchor` is snapped, >> but it probably should be. > > Yeah I think so too. I still think this code is incorrect. To illustrate, let's set insets to 0.8 and anchor to 0.5 Here is what I think it should do: snap=off result=1.3 snap=on result=1.0 (and not 1.5, as I think the proposed fix would produce) ------------- PR: https://git.openjdk.org/jfx/pull/910
