On Fri, 3 Jun 2022 05:42:38 GMT, Tejesh R <t...@openjdk.org> wrote: > I see that the bug description has a test case which produce this exception. > There are these lines on the stack: > > ``` > at > sun.awt.SunGraphicsCallback.runComponents(SunGraphicsCallback.java:116) > at java.awt.Container.paint(Container.java:1978) > at java.awt.Window.paint(Window.java:3906) > ``` > > So your automatic test may override the paint method and catch an exception > there. The main method will check some flags at the end.
@TejeshR13's latest version installs `UncaughtExceptionHandler` to see whether any exception occurs on EDT. > Can we iterate all installed L&F? It seems the test does not have any windows > specific code. It's Windows-specific because NPE is thrown on Windows only. However, testing other L&Fs makes sense to ensure it doesn't happen there either. > I think that the rendering to the BufferedImage as suggested by @aivanov-jdk > should work as well. In theory only. I was unable to make the table paint with its header to `BufferedImage`. The component isn't laid out until it's added to peered top-level. If you know how to make it paint as if in the frame, I'd welcome it. At this point, painting to `BufferedImage` doesn't reproduce the NPE. ------------- PR: https://git.openjdk.java.net/jdk/pull/8830
