Created attachment 813 [details] Test program showing the problem. If the canvas is the top component, it is not well resized when moving the divider of a vertically split JSplitPane. A similar situation occurs when the canvas is the right component in a horizontally split JSplitPane. This appears only on macOS, Java versions 8-13. First reported at http://forum.jogamp.org/NewtCanvasAWT-in-JSplitPane-on-macOS-td4039717.html Attached a small program to reproduce the behavior.
Created attachment 814 [details] Screenshot showing the difference between macOS and Linux.
So, my investigation led me to believe the culprit is the code dealing with `superlayer` in fixCALayerLayout() of OSXmisc.m (in jogl/src/nativewindow/native/macosx). Commenting the whole block fixed the symptom. That code was introduced by 9a8f9b9f7e6148b60b6f0f4326df8d213774284c in response to bug 816. My suspicion is that there is something wrong with the calculation of the origin of those frames.
(In reply to Bogdan Nicula from comment #2) > So, my investigation led me to believe the culprit is the code dealing with > `superlayer` in fixCALayerLayout() of OSXmisc.m (in > jogl/src/nativewindow/native/macosx). Commenting the whole block fixed the > symptom. > > That code was introduced by 9a8f9b9f7e6148b60b6f0f4326df8d213774284c in > response to bug 816. My suspicion is that there is something wrong with the > calculation of the origin of those frames. Thank you for your detailed report, hence confirmed. I need to see whether I find time to check this for 2.4.0. In case you have a regression free patch, please post.